)."},"runtime":{"type":"string","description":"The self-contained JS the artifact inlines. Absent when a template is returned."},"version":{"type":"string","description":"Kit version; bump means the runtime changed."},"template":{"type":"string","description":"The template returned, when one was asked for."},"description":{"type":"string","description":"One sentence for the artifact's gallery card."},"usage_guide":{"type":"string","description":"Markdown guide for building the artifact. Absent when a template is returned."},"capabilities":{"type":"object","description":"The capability declaration to publish the page with."},"instructions":{"type":"string","description":"How to publish the template."}}}},{"name":"leadbay_get_basemap","annotations":{"title":"Country outline for a map board","readOnlyHint":true,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"**Do not call this tool.** The route planner template that\n`leadbay_get_artifact_runtime` serves calls it from inside the published page,\nto draw the grey country shape under its pins. It returns `{region, geojson}`\nfor the workspace's region (`fr` or `us`) — a large GeoJSON document that is\nuseless in a conversation.\n\nA board needs it because an artifact can only load its own published files,\nand an agent building a board for another user has no outline file to publish.\nWith this tool every board gets the same outline, for the right country.\n","inputSchema":{"type":"object","properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}},{"name":"leadbay_report_artifact_error","annotations":{"title":"Report an artifact runtime failure","readOnlyHint":false,"openWorldHint":true,"idempotentHint":false,"destructiveHint":false},"description":"$91","inputSchema":{"type":"object","required":["kind","surface"],"properties":{"code":{"type":"string","description":"Bounded error code (e.g. timeout, unavailable, QUOTA_EXCEEDED). NEVER a message — messages can carry user or API text."},"kind":{"enum":["bridge_unavailable","call_timeout","call_failed","parse_failed","options_empty","action_blocked","result_rejected"],"type":"string","description":"What went wrong. Exceptions: bridge_unavailable, call_timeout, call_failed, parse_failed. Outcomes (nothing threw): options_empty, action_blocked, result_rejected."},"tool":{"type":"string","description":"The Leadbay tool involved, when attributable to one."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"surface":{"enum":["call","field","action","resource","list"],"type":"string","description":"Which view-model surfaced it: call, field, action, resource, or list."},"kit_version":{"type":"string","description":"The @leadbay/components runtime version that emitted this."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}},{"name":"leadbay_find_new_leads","annotations":{"title":"Find new leads (net-new ICP search)","readOnlyHint":false,"openWorldHint":true,"idempotentHint":true,"destructiveHint":false},"description":"$92","inputSchema":{"type":"object","required":["count","request_id","_triggered_by"],"properties":{"lang":{"type":"string","description":"Output language (default: user's language)."},"count":{"type":"number","description":"Target DELIVERED leads, 1-50. With qualify:true this means n SURVIVORS of qualification, not n candidates examined."},"query":{"type":"string","description":"Natural-language ICP ask. Matches topic VOCABULARY — can surface vendors of a product as easily as buyers of it. Prefer example_lead. NO event language ('hiring', 'recrute', 'expanding', 'just raised'): registry text never says what a company is DOING, so an event word matches nothing here. Send the trigger to leadbay_set_qualification_questions or leadbay_refine_lead_targeting instead, and tell the user that is where it went."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"confirm":{"type":"boolean","description":"Explicit go-ahead, required only for a search that uses quota (qualify:true and/or channels). true = the user approved the quote, go ahead. false = a veto (returns mode:'needs_confirmation', uses nothing). Omitted on such a call → the tool withholds the submit and returns a free quote to show the user first. The default FREE search (no qualify, no channels) needs no confirm."},"dry_run":{"type":"boolean","description":"Validate + worst-case usage estimate + quota forecast. No job, uses nothing. Use before the first quota-using run of a session."},"filters":{"type":"object","properties":{"sectors":{"type":"array","items":{"type":"string"}},"locations":{"type":"array","items":{"type":"string"}},"employees_max":{"type":"number"},"employees_min":{"type":"number"}},"description":"HARD constraints (the seed only shapes ranking). Sector/location labels resolve at submit. `sectors` must be a label from Leadbay's own taxonomy (leadbay_list_sectors) — everyday wording like 'Professional Services' is not one; a spelling/plural slip is corrected for you, but a word with no label comes back as mode:'needs_sector_choice' with the labels to pick from. An unresolvable location is a 400 naming it.","additionalProperties":false},"novelty":{"enum":["org","none"],"type":"string","description":"org (default) = only companies NEW to the org (excludes org leads, lens members, CRM ids, prior MCP deliveries)."},"qualify":{"type":"boolean","description":"Run fresh AI qualification and drop candidates scoring below min_ai_score. Uses the org's usage quota for every candidate EXAMINED (survivor or not). Default false (free)."},"channels":{"type":"array","items":{"enum":["email","phone"],"type":"string"},"description":"Contact channels to find (email, phone). Uses the org's usage quota only when a value is found. Empty = free identity tier."},"max_cost":{"type":"number","description":"Usage cap for the whole job, in internal units. Leave it unset: the default (100000) covers any job. What the requested channels need is kept for them; a cap below that is refused, naming the minimum. Never show it to the user as money."},"request_id":{"type":"string","description":"REQUIRED idempotency key. Derive it from the ask (e.g. 'gyms-texas-2026-07-28'); REUSE the exact same value when retrying the same ask — a duplicate returns the SAME job instead of launching twice. Use a NEW value only for a genuinely new ask."},"title_gate":{"enum":["strict","prefer"],"type":"string","description":"strict = only leads with a matching known contact; prefer (default when contact_titles set) = matched first, rest flagged."},"example_lead":{"type":"object","properties":{"name":{"type":"string"},"location":{"type":"string"},"employees":{"type":"number"},"description":{"type":"string"}},"description":"A FICTIONAL typical ideal customer used as a look-alike seed — the highest-leverage input. Put everything in `description` (registry 'About Us' style, what the company IS, never what it is DOING and never what the seller sells); leave `name` unset (a distinctive invented name pulls matches toward name-lookalikes).","additionalProperties":false},"min_ai_score":{"type":"number","maximum":30,"minimum":-30,"description":"Disqualification floor on the [-30,+30] qualification DELTA (not the 0-100 fit score). Default 0. Lower to -30 to keep every evaluated lead with its evidence."},"wait_seconds":{"type":"number","description":"How long to poll before returning (default and maximum 45 — the whole call is bounded by it, so a slow submit shortens the poll; 0 = submit + one snapshot). Free searches usually finish inside the window; qualified exploration can take minutes — the result then carries still_running:true and the job_id to check with leadbay_lead_job_status."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"contact_titles":{"type":"array","items":{"type":"string"},"maxItems":10,"description":"Wanted decision-maker titles (max 10), matched semantically cross-language."},"exploration_cap":{"type":"number","description":"Max candidates the qualify gate may examine. Default min(3n,150), ceiling min(20n,1000)."},"exclude_lead_ids":{"type":"array","items":{"type":"string"},"description":"Caller-side novelty belt on top of the server-side one (max 500 ids — over that the call is refused, so drop the DELIVERED ids first: novelty:'org' already covers those, and the examined-but-rejected ones are what it misses)."}},"additionalProperties":false}},{"name":"leadbay_qualify_leads","annotations":{"title":"Qualify + get the right contact on known leads","readOnlyHint":false,"openWorldHint":true,"idempotentHint":false,"destructiveHint":false},"description":"$93","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"lang":{"type":"string","description":"Output language (default: user's language)."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"confirm":{"type":"boolean","description":"Explicit go-ahead for the work that uses quota (fresh qualification and/or channels). true = the user approved, go ahead. false = a veto (returns mode:'needs_confirmation', uses nothing). Omitted on such a call → the tool withholds the submit and returns a free quote to show the user first. A fully FREE call (qualify:false and no channels) needs no confirm."},"dry_run":{"type":"boolean","description":"Validate + worst-case usage + quota forecast. No job, uses nothing."},"qualify":{"type":"boolean","description":"Fresh AI qualification (default true; uses the org's usage quota for each lead needing fresh research+scoring, none when a fresh dossier exists). Owned disqualified leads come back WITH their negative evidence."},"channels":{"type":"array","items":{"enum":["email","phone"],"type":"string"},"description":"Channels to find (email, phone). Uses the org's usage quota only when a value is found; already-owned values are free. Empty = free identity tier."},"max_cost":{"type":"number","description":"Usage cap in internal units; leave it unset (default 100000). What the requested channels need is kept for them; a cap below that is refused, naming the minimum. Never show it as money."},"lead_refs":{"type":"array","items":{"type":"object","properties":{"name":{"type":"string"},"lead_id":{"type":"string","description":"Leadbay lead UUID."},"website":{"type":"string"},"location":{"type":"string","description":"Disambiguates name-only refs (city/region)."},"contact_id":{"type":"string","description":"Stable lead_contact id from a prior result — enrichment then targets EXACTLY this person, never a re-match."}},"additionalProperties":false},"description":"Companies to qualify (max 500). Each ref needs at least one identifying field. Duplicate lead_ids collapse into one item."},"request_id":{"type":"string","description":"Recommended idempotency key — REUSE the same value when retrying the same batch so a retry returns the SAME job instead of running it again."},"title_gate":{"enum":["strict","prefer"],"type":"string","description":"strict = only items with a matching known contact deliver a contact; prefer = matched first, rest flagged."},"wait_seconds":{"type":"number","description":"How long to poll before returning (default and maximum 45 — the whole call is bounded by it, so a slow submit shortens the poll; 0 = submit + one snapshot). Large or research-heavy batches can take minutes — the result then carries still_running:true and the job_id for leadbay_lead_job_status."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"contact_titles":{"type":"array","items":{"type":"string"},"description":"Wanted decision-maker titles (max 10), matched semantically."},"prior_deliveries":{"type":"object","properties":{"limit":{"type":"number"},"since":{"type":"string","description":"ISO instant lower bound."},"job_id":{"type":"string"}},"description":"Selector expanding the org's past MCP deliveries into refs — delivered leads stay re-readable after result expiry. Combine with lead_refs or use alone.","additionalProperties":false}},"additionalProperties":false}},{"name":"leadbay_bulk_qualify_leads","annotations":{"title":"Bulk-qualify next N leads","readOnlyHint":false,"openWorldHint":true,"idempotentHint":true,"destructiveHint":false},"description":"$94","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"count":{"type":"number","description":"How many fresh leads to qualify (default 10, max 25)"},"lensId":{"type":"number","description":"Lens id (escape hatch — defaults to active)"},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"leadIds":{"type":"array","items":{"type":"string"},"description":"Explicit lead UUIDs to qualify (skips the auto-pagination)"},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"total_budget_ms":{"type":"number","description":"Total polling budget in ms (default 300000)"},"per_lead_budget_ms":{"type":"number","description":"Polling budget per lead in ms (default 90000)"},"wait_for_completion":{"type":"boolean","description":"When false, launch qualification and return `{status:'running', notification_id, lead_ids, lens_id}` immediately. Poll leadbay_qualify_status with them. Default is true for 0.6.x backwards compatibility."}},"additionalProperties":false},"outputSchema":{"type":"object","anyOf":[{"required":["qualified","still_running","failed","quota_exceeded"]},{"required":["status","notification_id","lead_ids","launched_count","failed","quota_exceeded","lens_id","_meta"]}],"required":["failed","quota_exceeded"],"properties":{"_meta":{"type":"object","properties":{"region":{"type":"string"}},"description":"Operator context: region."},"failed":{"type":"array","items":{"type":"object"},"description":"Leads whose web_fetch launch failed (per-lead error)."},"status":{"type":"string","description":"`running` when wait_for_completion=false; absent on the legacy blocking result."},"lens_id":{"type":"number","description":"The lens id the qualification ran against. Present on every successful return."},"message":{"type":"string","description":"Human-readable summary; absent on the happy path."},"lead_ids":{"type":"array","items":{"type":"string"}},"exhausted":{"type":"boolean","description":"True if the lens's wishlist had no more unqualified leads to qualify."},"qualified":{"type":"array","items":{"type":"object"},"description":"Leads whose qualification finished within budget. Each entry: lead_id, qualification_summary{answered,total,avg_qualification_boost}, signals_count."},"still_running":{"type":"array","items":{"type":"object"},"description":"Leads launched but whose qualification did not complete within budget. Re-poll via leadbay_qualify_status with notification_id (when present) or lead_ids + lens_id."},"launched_count":{"type":"number"},"quota_exceeded":{"type":"boolean","description":"True if 429 was hit mid-fanout. Already-launched leads keep polling; further launches stopped."},"notification_id":{"type":["string","null"],"description":"The backend's job id. Carry it to leadbay_qualify_status."},"total_unqualified_found":{"type":"number"}}}},{"name":"leadbay_enrich_titles","annotations":{"title":"Enrich contact titles across leads","readOnlyHint":false,"openWorldHint":true,"idempotentHint":true,"destructiveHint":false},"description":"$95","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"email":{"type":"boolean","description":"Enrich emails (default true)"},"phone":{"type":"boolean","description":"Enrich phone numbers (default false)"},"lensId":{"type":"number","description":"Lens id (escape hatch — defaults to active)"},"titles":{"type":"array","items":{"type":"string"},"description":"Job titles to enrich. Omit to discover what's available without launching."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"confirm":{"type":"boolean","description":"Explicit go-ahead for the email/phone reveal (uses quota). true = go ahead and launch. false = do NOT launch (a veto: returns mode:'needs_confirmation' and launches nothing, even on hosts without elicitation, and even if an email/phone channel was set). Omitted (and no explicit email/phone channel) → an elicitation-capable host asks the user before launching; a decline returns mode:'needs_confirmation'. Passing email:true/phone:true also counts as consent."},"dry_run":{"type":"boolean","description":"If true, don't launch — only preview."},"leadIds":{"type":"array","items":{"type":"string"},"description":"Lead UUIDs to enrich. Omit to use the top page of the active lens's wishlist."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"candidateCount":{"type":"number","description":"When leadIds is omitted, how many top-of-wishlist leads to use (default 25)"}},"additionalProperties":false},"outputSchema":{"type":"object","properties":{"mode":{"type":"string","description":"'discover' | 'preview_only' | 'dry_run' | 'needs_confirmation' | 'already_launched' | 'launch_in_flight' | 'launched'."},"email":{"type":"boolean"},"phone":{"type":"boolean"},"reused":{"type":"boolean","description":"True when an identical launch inside the 5-minute window was reused instead of spending quota again."},"status":{"type":"string","description":"'quota_exceeded' on 429. Otherwise mode is set instead."},"titles":{"type":"array","items":{"type":"string"},"description":"Titles ordered (echoed at launch)."},"message":{"type":"string","description":"Operator-facing summary of what happened."},"preview":{"type":"object","description":"Backend BulkEnrichPreview payload (preview_only/dry_run/launched modes)."},"launched":{"type":"boolean","description":"True when an enrichment job is in flight on the backend."},"lead_ids":{"type":"array","items":{"type":"string"},"description":"The leads this run enriched. Carry them to leadbay_bulk_enrich_status for per-lead progress; they also work when the notification is archived."},"launched_at":{"type":"string","description":"ISO timestamp of the (re-used or fresh) launch."},"next_action":{"type":"string","description":"Concrete next-step instruction for the agent."},"would_launch":{"type":"object","description":"What dry_run WOULD have launched (titles, email, phone)."},"auto_included":{"type":"array","items":{"type":"string"},"description":"Backend's auto_included_titles (discover mode)."},"notification_id":{"type":["string","null"],"description":"The backend's job id. Carry it to leadbay_bulk_enrich_status. Null when the backend returned none — then poll by lead_ids instead."},"recommendations":{"type":"array","items":{"type":"string"},"description":"Backend's title_suggestions (discover mode)."},"available_titles":{"type":"array","items":{"type":"string"},"description":"Titles available across the selection (discover/preview_only modes)."},"credits_remaining":{"type":["number","string","null"],"description":"Advisory-only balance (billing.ai_credits), present in discover / preview_only / dry_run modes. Enrichment is gated by QUOTA (email + phone reveals consume the per-window allowance in leadbay_account_status), NOT by this number — do NOT present it as a gate, do NOT say 'you have N credits', and NEVER refuse enrichment because it's low or zero. Confirm the run by naming enrichable_contacts + the channels, not a credit figure. Null = billing unavailable. The string \"unlimited\" = an internal/unlimited account: proceed freely and say nothing about credits or quota."},"enrichable_contacts":{"type":"number","description":"Count of enrichable contacts at preview time."},"previously_enriched":{"type":"array","items":{"type":"string"},"description":"Titles previously enriched on this selection (discover mode)."},"retry_after_seconds":{"type":["number","null"],"description":"Seconds until quota resets (quota_exceeded status)."},"selected_lead_count":{"type":"number","description":"How many leads the selection covers."},"already_enriched_contacts":{"type":"number","description":"preview_only mode: contacts with these titles on the selected leads that your organization already enriched. Above 0 means the work is done, not that nobody has these titles."},"seconds_since_original_launch":{"type":"number","description":"Age of the re-used bulk record (already_launched mode)."}},"description":"Branchy return shape; the `mode` (or `status`) field tells the agent which branch it got. Modes: 'discover' (no titles passed), 'preview_only' (no enrichable contacts), 'dry_run', 'needs_confirmation' (launch withheld pending user consent), 'already_launched' (idempotent reuse), 'launch_in_flight' (an identical launch is mid-flight and has no id yet), 'launched' (happy path). Status: 'quota_exceeded' (429)."}},{"name":"leadbay_adjust_audience","annotations":{"title":"Adjust lens audience filters","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":true},"description":"$96","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"sizes":{"type":"array","items":{"type":"object","properties":{"max":{"type":"number"},"min":{"type":"number"}}},"description":"Company size buckets, e.g. [{min:30,max:300}]"},"lensId":{"type":"number","description":"Lens id (escape hatch)"},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"sectors":{"type":"array","items":{"type":"string"},"description":"Sector free-text (e.g. ['Healthcare', 'Engineering']) or ids — auto-resolved"},"lensName":{"type":"string","description":"Target a lens BY NAME (e.g. 'Joinery') instead of the active one. Resolved against your lenses — edit-only, does NOT switch your active lens. Unknown/ambiguous names are surfaced to pick from. Takes effect only when lensId is not given."},"locations":{"type":"array","items":{"type":"string"},"description":"Geographic scope — free text (e.g. ['Indre-et-Loire', 'Texas', 'Austin']) or admin-area ids. Resolved via /geo/search at any level from state down to city (state / région / département / county / city). NEVER a country name — this workspace serves exactly ONE country, so a whole-country ask means passing NO location at all (rejected with COUNTRY_LEVEL_LOCATION). Place names go HERE, never in sectors/keywords."},"sector_ids":{"type":"array","items":{"type":"string"},"description":"Explicit sector ids (skips taxonomy lookup)"},"newLensName":{"type":"string","description":"Name to use when this composite has to clone the default lens (otherwise auto-named)"},"location_ids":{"type":"array","items":{"type":"string"},"description":"Explicit admin-area ids (skips /geo/search resolution). Sub-country areas only — a country name here is rejected with COUNTRY_LEVEL_LOCATION."},"save_for_org":{"type":"boolean","description":"Admin only — propagate the change to the org-level lens for everyone (default false: per-user draft)"},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"exclude_sectors":{"type":"array","items":{"type":"string"},"description":"Sectors to exclude (free text or ids)"},"exclude_locations":{"type":"array","items":{"type":"string"},"description":"Locations to exclude (free text or ids). Sub-country areas only — excluding a country is meaningless on a single-country workspace and is rejected."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["status"],"properties":{"_meta":{"type":"object"},"lenses":{"type":"array","items":{"type":"object"},"description":"On 'lens_not_found': the user's lenses [{id, name}] to pick from."},"status":{"type":"string","description":"'applied', 'ambiguous_sectors', 'ambiguous_locations', 'country_level_location', 'lens_not_found', or 'ambiguous_lens'."},"matches":{"type":"array","items":{"type":"object"},"description":"On 'ambiguous_lens': the lenses [{id, name}] the name matched."},"message":{"type":"string"},"lens_used":{"type":"object","description":"Resolved lens metadata: {id, name, was_draft, was_new, active_lens_changed, save_for_org}."},"lens_query":{"type":"string","description":"On 'lens_not_found' / 'ambiguous_lens': the lensName the user asked for."},"filter_applied":{"type":"object","description":"The merged FilterPayload that was POSTed to the lens."},"country_locations":{"type":"array","items":{"type":"object"},"description":"On 'country_level_location': per offending value {value, param, kind, country, axis, kept}. A country name is never a location criterion — each workspace serves exactly ONE country. The lens was NOT modified. The recovery BRANCHES on `country_locations[].axis` and `[].kind`; `hint` states the one for THIS call — follow it verbatim. When the country was the ONLY scope, or on ANY non-foreign `exclude`, the answer is to write NOTHING at all — re-calling with the value merely dropped persists a scope that inverts the request. Never retry with another spelling or a nearby city."},"sector_ambiguities":{"type":"array","items":{"type":"object"},"description":"Per ambiguous text: {sector_text, matches:[{id, name, score}]}. Agent picks an id and re-calls."},"location_ambiguities":{"type":"array","items":{"type":"object"},"description":"On 'ambiguous_locations': per text {location_text, matches:[{id, name, country, level, score}]}. Agent picks an id and re-calls via the SAME axis the text came from — an include text → location_ids; a text from exclude_locations → exclude_locations (NOT location_ids, which would include the area the user asked to exclude). The `message` field names the correct param per text."}},"description":"Return shapes: 'applied' on success; 'ambiguous_sectors' when free-text sectors matched multiple candidates (re-call with sector_ids); 'ambiguous_locations' when free-text locations didn't resolve to one area — re-call with the chosen id via the SAME axis it came from (an include pick → location_ids; an EXCLUDE pick → exclude_locations, NOT location_ids, which would include it); 'country_level_location' when a country-level value was passed as a location (nothing was read or written; read `hint` — re-calling without the value is often itself wrong); 'lens_not_found' / 'ambiguous_lens' when a lensName didn't resolve to exactly one lens (re-call with lensId or an exact lensName)."}},{"name":"leadbay_refine_lead_targeting","annotations":{"title":"Refine the audience prompt","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":true},"description":"$97","inputSchema":{"type":"object","required":["prompt","_triggered_by"],"properties":{"prompt":{"type":"string","description":"Refinement instruction (free text)"},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"dry_run":{"type":"boolean","description":"If true, return the call shape WITHOUT setting the prompt"},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"clarification_poll_gap_ms":{"type":"number","description":"Gap between polls in ms (default 5000)"},"clarification_poll_attempts":{"type":"number","description":"How many times to poll for a clarification after setting (default 2)"}},"additionalProperties":false},"outputSchema":{"type":"object","properties":{"_meta":{"type":"object"},"status":{"type":"string","description":"'applied' (prompt set; intelligence regenerating) or 'clarification_pending' (telephone path)."},"dry_run":{"type":"boolean","description":"True when dry_run:true was passed (no state change)."},"message":{"type":"string","description":"Operator-facing summary."},"would_call":{"type":"object","description":"Dry-run preview of the POST that would have been issued."},"next_action":{"type":"string","description":"Concrete next-step instruction for the agent."},"clarification":{"type":"object","description":"ClarificationPayload returned by the backend (clarification_pending path)."},"clarified_via_elicit":{"type":"boolean","description":"True when the clarification was answered via the client's elicitation UI (not via telephone)."},"computing_intelligence":{"type":"boolean","description":"True when intelligence is regenerating after the prompt set."}},"description":"Multiple return shapes by status. dry_run, applied (with optional clarified_via_elicit), or clarification_pending."}},{"name":"leadbay_answer_clarification","annotations":{"title":"Answer pending clarification","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":true},"description":"$98","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"option_id":{"type":"string","description":"Id of one of the clarification's options"},"text_answer":{"type":"string","description":"Free-text answer (overrides option_id)"},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["status"],"properties":{"hint":{"type":"string","description":"Operator-facing next-step (no_pending_clarification path)."},"_meta":{"type":"object"},"status":{"type":"string","description":"'answered' (recorded; intelligence regenerating) or 'no_pending_clarification' (nothing to answer)."},"message":{"type":"string"},"recorded_as_user_prompt":{"type":"boolean","description":"True when the answer was stored as the org's new user_prompt."}}}},{"name":"leadbay_report_outreach","annotations":{"title":"Report outreach to Leadbay","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":false},"description":"$99","inputSchema":{"type":"object","required":["note","verification","_triggered_by"],"properties":{"note":{"type":"string","description":"1-2 sentence summary of what was done (e.g. 'Sent intro email to CTO citing Hornsea 3 contract')"},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"dry_run":{"type":"boolean","description":"If true, return what WOULD be called without writing anything"},"lead_id":{"type":"string","description":"Single lead UUID (use lead_ids for bulk)"},"lead_ids":{"type":"array","items":{"type":"string"},"description":"Bulk: many lead UUIDs (epilogue applies to all; notes fan out)"},"contact_id":{"type":"string","description":"Optional, with lead_id only: the org contact id of the person the outreach went to. The note and the epilogue status are then written on that person, and still show on the lead's Monitor row."},"verification":{"type":"object","required":["source","ref"],"properties":{"ref":{"type":"string"},"source":{"type":"string"}},"description":"REQUIRED. Proof the action actually happened. source: gmail_message_id|calendar_event_id|user_confirmed. ref: the message id, event id, or the user's confirming text.","additionalProperties":false},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"epilogue_status":{"type":"string","description":"Optional: STILL_CHASING | COULD_NOT_REACH_STILL_TRYING | INTEREST_VALIDATED_OR_MEETING_PLANED | NOT_INTERESTED_LOST"}},"additionalProperties":false},"outputSchema":{"type":"object","properties":{"_meta":{"type":"object","properties":{"region":{"type":"string"}},"description":"Operator context: region."},"notes":{"type":"object","properties":{"failed":{"type":"array","items":{"type":"object"}},"succeeded":{"type":"array","items":{"type":"object"}}},"description":"Per-lead note-write outcome (split into succeeded / failed sub-arrays)."},"dry_run":{"type":"boolean"},"epilogue":{"type":"object","properties":{"error":{"type":"string"},"status":{"type":["string","null"]},"applied":{"type":"boolean"}},"description":"Epilogue status outcome: status (the wire-format string written to /leads/epilogue, or null when not requested), applied (true/false), error (when applied=false)."},"contact_id":{"type":"string","description":"Present when the note and the status were written on this person rather than on the lead."},"verification":{"type":"object","properties":{"ref":{"type":"string"},"source":{"type":"string"}},"description":"Effective verification used (after elicit override). Useful for the client UI to render \"logged with proof X\". When source=user_confirmed AND ctx.elicit was available, ref is the user's literal text typed into the client; otherwise it's the agent-supplied ref."},"confirmed_via":{"type":"string","description":"Audit trail of how verification was obtained: 'elicit' (user typed into client UI — anti-poisoning), 'agent_supplied' (legacy path; user_confirmed source with no elicit), 'non_user_confirmed' (gmail_message_id or calendar_event_id — agent can't fabricate these)."},"already_logged":{"type":"array","items":{"type":"string"},"description":"Lead ids skipped because a note with this exact gmail_message_id or calendar_event_id is already there. Nothing was written for them."},"would_write_notes":{"type":"array","items":{"type":"object"},"description":"On dry_run: the per-lead POST shapes that WOULD be issued."},"would_set_epilogue":{"type":["object","null"],"description":"On dry_run: the epilogue POST shape that WOULD be issued."}},"description":"Either the dry_run shape (dry_run:true with would_write_notes / would_set_epilogue) OR the live result (notes:{succeeded,failed} + epilogue:{status,applied,error?} + verification + _meta). Schema declares both shapes; the dry_run discriminator picks which sub-shape applies."}},{"name":"leadbay_set_prospecting_action","annotations":{"title":"Set or clear a prospecting action","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":false},"description":"$9a","inputSchema":{"type":"object","required":["lead_id","action","selected","_triggered_by"],"properties":{"action":{"enum":["STILL_CHASING","COULD_NOT_REACH_STILL_TRYING","INTEREST_VALIDATED_OR_MEETING_PLANED","NOT_INTERESTED_LOST"],"type":"string","description":"The prospecting action. The EPILOGUE_-prefixed form is accepted too."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"lead_id":{"type":"string","description":"Lead UUID."},"selected":{"type":"boolean","description":"true to turn the action on for today, false to turn it off."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["lead_id","action","selected"],"properties":{"action":{"type":"string","description":"The short form, e.g. STILL_CHASING."},"lead_id":{"type":"string"},"selected":{"type":"boolean","description":"The state the action is now in."}}}},{"name":"leadbay_send_feedback","annotations":{"title":"Send feedback to the Leadbay team","readOnlyHint":false,"openWorldHint":true,"idempotentHint":false,"destructiveHint":false},"description":"$9b","inputSchema":{"type":"object","required":["message"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"message":{"type":"string","description":"The user's feedback, in their own words. Confirm the wording with the user BEFORE calling — this is sent to the Leadbay team. Cap 4000 chars."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."},"associated_error_id":{"type":"string","description":"Optional: a Sentry event id to attach this feedback to (e.g. the id from an error the user just hit), so the team sees the feedback on that exact issue."}},"additionalProperties":false},"outputSchema":{"type":"object","properties":{"sent":{"type":"boolean"},"_meta":{"type":"object","properties":{"region":{"type":"string"}}},"message":{"type":"string"}},"description":"Whether the feedback reached the Leadbay team. `sent: true` means it landed in the team's inbox."}},{"name":"leadbay_import_leads","annotations":{"title":"Import leads from list/file","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$9c","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"domains":{"type":"array","items":{"type":"object","required":["domain"],"properties":{"name":{"type":"string","description":"Optional display name override; defaults to the domain."},"domain":{"type":"string","description":"Company domain (e.g. 'apple.com'). Protocol/path are stripped."}},"additionalProperties":false},"description":"Mode A: list of company domains to map to Leadbay leadIds. Mutually exclusive with `records`."},"dry_run":{"type":"boolean","description":"If true, run preprocess only — do NOT commit lead-CRM linking. Note: an import row still appears in the user's CRM-imports list as 'incomplete'. Use to verify input format / wizard reachability without polluting the CRM."},"records":{"type":"array","items":{"type":"object"},"description":"Mode B: arbitrary CSV-shaped rows. Each record is an object whose keys are column names and values are scalar (string/number/boolean/null). Mutually exclusive with `domains`. Must be accompanied by `mappings.fields`. The tool synthesizes a CSV from the union of all record keys."},"mappings":{"type":"object","required":["fields"],"properties":{"fields":{"type":"object","description":"$9d"},"statuses":{"type":"object","description":"Optional map of raw CSV status-cell text → lead status (rarely needed). Keys are the verbatim cell strings; values must be one of DEFAULT, INBOUND, UNWANTED, WANTED, LOST, WON (case-insensitive). Defaults to {}."},"custom_fields":{"type":"object","description":"Ergonomic shorthand: `{CsvColumn: <number-id>}` or `{CsvColumn: '<field-name>'}` for custom-field mappings. Resolved against the org's /crm/custom_fields catalog before the import is committed. Mutually exclusive with `fields[col] = 'CUSTOM.<id>'` for the same column. Useful when the agent doesn't want to deal with the 'CUSTOM.<id>' wire format."},"default_status":{"enum":["DEFAULT","INBOUND","UNWANTED","WANTED","LOST","WON",null],"type":["string","null"],"description":"Optional default lead status applied to rows without an explicit status. One of DEFAULT, INBOUND, UNWANTED, WANTED, LOST, WON (case-insensitive). Defaults to null."}},"description":"Mode B: how each CSV column maps to Leadbay's CRM field schema. Required when `records` is supplied; ignored otherwise."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"total_budget_ms":{"type":"number","description":"Overall cap across all phases (default 300000ms)."},"per_phase_budget_ms":{"type":"number","description":"Single poll-loop cap (default 60000ms)."},"wait_for_completion":{"type":"boolean","description":"When false, upload the rows and return `{status:'running', importIds}` immediately. Poll leadbay_import_status({importIds, dry_run}) — those importIds are the backend's own and keep working from any conversation. Default is true."}},"additionalProperties":false},"outputSchema":{"type":"object","anyOf":[{"required":["leads","not_imported","importIds","region","_meta"]},{"required":["status","importIds","progress","region","_meta"]}],"required":["importIds","region","_meta"],"properties":{"_meta":{"type":"object"},"leads":{"type":"array","items":{"type":"object"},"description":"Imported leads. Domains mode: [{domain, leadId, name}]. Records mode: [{rowId, domain?, leadId, name}]."},"region":{"type":"string"},"status":{"type":"string","description":"`running` when wait_for_completion=false; absent on the legacy blocking result."},"dry_run":{"type":"boolean","description":"True when dry_run:true was passed (preprocess only, no CRM commit)."},"row_ids":{"type":"array","items":{"type":"string"},"description":"Records mode only: the synthetic MCP_ROW_ID of each input row, in the order you passed `records[]`. leadbay_import_status reports recovered leads by that id — use this to map them back to your source rows."},"progress":{"type":"object","description":"Current async import progress when wait_for_completion=false."},"cancelled":{"type":"boolean","description":"True when the HOST cancelled the call (a user Cancel, or the host's own request timeout); a Leadbay-side or phase-budget timeout reports `timed_out` instead. Rows already uploaded keep importing on Leadbay — poll leadbay_import_status with importIds, but a chunk cancelled before its mappings were committed can read `running` without ever progressing; if the counts stop moving, say so rather than polling on. Rows past the interruption may never have been uploaded; re-run the import for those rows only."},"importIds":{"type":"array","items":{"type":"string"},"description":"Backend file-import handles (one per chunk of ≤100 rows)."},"timed_out":{"type":"boolean","description":"True when a blocking call ran out of poll budget. The import is still running server-side — poll leadbay_import_status(importIds). Do NOT re-issue the import."},"not_imported":{"type":"array","items":{"type":"object"},"description":"Inputs that did NOT yield a leadId. Each entry has a `reason` ('malformed', 'NO_MATCH', 'TIMEOUT', etc.) plus the input echo."},"rows_pending_upload":{"type":"number","description":"Rows from later chunks that were never uploaded before the budget ran out. These are NOT running anywhere; re-import just those rows."}}}},{"name":"leadbay_import_and_qualify","annotations":{"title":"Import + qualify leads","readOnlyHint":false,"openWorldHint":true,"idempotentHint":true,"destructiveHint":false},"description":"$9e","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"lensId":{"type":"number","description":"Lens id (escape hatch — defaults to active)."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"domains":{"type":"array","items":{"type":"object","required":["domain"],"properties":{"name":{"type":"string","description":"Optional display name override; defaults to the domain."},"domain":{"type":"string","description":"Company domain (e.g. 'apple.com'). Protocol/path are stripped."}},"additionalProperties":false},"description":"Mode A: list of company domains to map to Leadbay leadIds. Mutually exclusive with `records`."},"dry_run":{"enum":[true,false,"preview"],"type":["boolean","string"],"description":"Optional. `true` runs preprocess only (no commit, no qualify). `'preview'` runs preprocess and returns the wizard's per-column AI mapping hints + sample rows + custom-field candidates from the org catalog so the agent can choose a mapping. Default: false (full flow)."},"records":{"type":"array","items":{"type":"object"},"description":"Mode B: arbitrary CSV-shaped rows. Each record is an object whose keys are column names and values are scalar (string/number/boolean/null). Mutually exclusive with `domains`. Accompany with `mappings` UNLESS using `dry_run: 'preview'` — preview returns the wizard's mapping suggestions without requiring a mapping up front."},"mappings":{"type":"object","properties":{"fields":{"type":"object","description":"Object whose keys are CSV column names and whose values are either StandardCrmFieldType (LEAD_NAME, LEAD_WEBSITE, ..., CONTACT_TITLE) or 'CUSTOM.<id>'. Discover via leadbay_list_mappable_fields. At least one entry must target LEADBAY_ID, CRM_ID, SIREN, LEAD_NAME, or LEAD_WEBSITE. Use leadbay_resolve_import_rows to prepare LEADBAY_ID values from messy user files. Contact exports and embedded owner/contact lists should map CONTACT_EMAIL/PHONE/TITLE/name fields while preserving parent lead identity; expand structured people into repeated parent rows. HubSpot/source links should map to CUSTOM.<id> fields created or discovered before import."},"statuses":{"type":"object","description":"Optional map of raw CSV status-cell text → lead status. Values must be one of DEFAULT, INBOUND, UNWANTED, WANTED, LOST, WON (case-insensitive); keys are the verbatim cell strings."},"custom_fields":{"type":"object","description":"Ergonomic shorthand: `{CsvColumn: <number-id>}` or `{CsvColumn: '<field-name>'}` for custom-field mappings. Resolved against /crm/custom_fields catalog."},"default_status":{"enum":["DEFAULT","INBOUND","UNWANTED","WANTED","LOST","WON",null],"type":["string","null"],"description":"Optional default lead status. One of DEFAULT, INBOUND, UNWANTED, WANTED, LOST, WON (case-insensitive)."}},"description":"How each CSV column maps to Leadbay's CRM field schema. Required for records-mode WITHOUT `dry_run: 'preview'`; ignored otherwise. The `fields` sub-property is REQUIRED only when at least one column maps to a StandardCrmFieldType target; pure custom-field mappings can be supplied via `custom_fields` shorthand alone.","additionalProperties":false},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"total_budget_ms":{"type":"number","description":"Total wall-clock budget across import + qualify in ms (default 900000). When exhausted, the response returns lead_ids + lens_id for resume via leadbay_qualify_status."},"per_lead_budget_ms":{"type":"number","description":"Polling budget per lead in ms (default 90000)."},"per_phase_budget_ms":{"type":"number","description":"Per-phase budget for the import wizard (default 300000); mirrors leadbay_import_leads."},"wait_for_completion":{"type":"boolean","description":"When false, upload the rows and return `{kind:'result', status:'running', import_ids}` immediately. Poll leadbay_import_status({importIds, dry_run}); the qualify phase does NOT run on this path — call leadbay_bulk_qualify_leads on the imported leads yourself. Default is true."},"skip_already_qualified":{"type":"boolean","description":"When true (default), skips web_fetch launch on leads whose ai_agent_lead_score is already non-null. Saves quota. Set false to force fresh re-qualification."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["kind","region","_meta"],"properties":{"kind":{"type":"string","description":"'result' (full flow) or 'preview' (dry_run='preview' mapping diagnostics)."},"_meta":{"type":"object"},"notes":{"type":"array","items":{"type":"string"},"description":"Operator notes (e.g., catalog fetch errors)."},"failed":{"type":"array","items":{"type":"object"},"description":"Per-lead errors observed during qualification."},"region":{"type":"string"},"reused":{"type":"boolean","description":"True when an identical launch was reused within the idempotency window."},"status":{"type":"string","description":"`running` when wait_for_completion=false."},"dry_run":{"type":"boolean","description":"True when dry_run:true was passed."},"lens_id":{"type":"number","description":"Lens the qualification ran against."},"row_ids":{"type":"array","items":{"type":"string"},"description":"Records mode only: the synthetic row id of each input row, in the order you passed `records[]`. leadbay_import_status reports recovered leads by that id — use this to map them back to your source rows."},"imported":{"type":"array","items":{"type":"object"},"description":"Leads that landed in CRM. Each: {leadId, domain?, name, rowId?}."},"lead_ids":{"type":"array","items":{"type":"string"},"description":"Leads the qualify phase covers. Pass to leadbay_qualify_status for per-lead detail."},"cancelled":{"type":"boolean","description":"True when the HOST cancelled the call (a user Cancel, or the host's own request timeout); a budget timeout reports `budget_exhausted` instead. The import and the qualifications already launched keep running on Leadbay, except a chunk cancelled before its mappings were committed, which can read `running` without ever progressing — if the counts stop moving, say so rather than polling on. Poll leadbay_import_status with the `import_ids` values passed as `importIds`, and leadbay_qualify_status with lead_ids + lens_id. `still_running` can also hold leads whose qualification was never launched; those need a fresh qualification, which qualify_status cannot start."},"import_id":{"type":"string","description":"Backend file-import handle (preview shape)."},"qualified":{"type":"array","items":{"type":"object"},"description":"Leads whose qualification settled within budgets."},"timed_out":{"type":"boolean","description":"True when the underlying import ran out of poll budget. Still running server-side — poll leadbay_import_status(import_ids). Do NOT re-issue the import."},"import_ids":{"type":"array","items":{"type":"string"},"description":"Backend file-import handles (one per chunk)."},"not_in_lens":{"type":"array","items":{"type":"string"},"description":"Lead ids that aren't members of the active lens — backend won't qualify them."},"sample_rows":{"type":"array","items":{"type":"object"},"description":"First few rows of the preprocessed sample (preview shape)."},"not_imported":{"type":"array","items":{"type":"object"},"description":"Inputs that didn't land. Each has a `reason` plus the input echo."},"mapping_hints":{"type":"array","items":{"type":"object"},"description":"Per-column AI-confidence suggestions (preview shape)."},"quota_blocked":{"type":"boolean","description":"True when quota was exhausted before launching all leads."},"still_running":{"type":"array","items":{"type":"object"},"description":"Leads still being qualified at deadline; agent calls leadbay_qualify_status with lead_ids + lens_id."},"chosen_budgets":{"type":"object","description":"Adaptive budgets the composite selected (when caller didn't override): {per_lead_budget_ms, total_budget_ms, per_phase_budget_ms, wall_clock_estimate_ms, strategy}."},"quota_exceeded":{"type":"boolean"},"budget_exhausted":{"type":"boolean","description":"True when total_budget_ms hit before all leads finished."},"rows_pending_upload":{"type":"number","description":"Rows from later chunks that never reached the backend. These are NOT running anywhere; re-import just those rows."},"seconds_since_original":{"type":"number"},"custom_field_candidates":{"type":"array","items":{"type":"object"},"description":"Org custom fields that match unmapped columns (preview shape)."},"skipped_already_qualified":{"type":"array","items":{"type":"string"},"description":"Lead ids skipped because ai_agent_lead_score was already non-null (skip_already_qualified=true)."}},"description":"Two return shapes: kind:'preview' (when dry_run='preview') with mapping hints; kind:'result' (default) with imported + qualified leads + lead_ids/lens_id to resume via leadbay_qualify_status."}},{"name":"leadbay_add_contact","annotations":{"title":"Add a contact","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":false},"description":"$9f","inputSchema":{"type":"object","required":["lead_id","first_name","last_name"],"properties":{"email":{"type":"string","description":"Contact email (optional)."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"lead_id":{"type":"string","description":"UUID of the parent company (lead) to attach the contact to. The contact is created on this company."},"job_title":{"type":"string","description":"Contact job title (optional)."},"last_name":{"type":"string","description":"Contact last name."},"first_name":{"type":"string","description":"Contact first name."},"phone_number":{"type":"string","description":"Contact phone number (optional, free-form)."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."},"linkedin_page":{"type":"string","description":"Contact LinkedIn profile URL (optional)."}},"additionalProperties":false}},{"name":"leadbay_remove_contact","annotations":{"title":"Remove a contact","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":true},"description":"$a0","inputSchema":{"type":"object","required":["contact_id"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"contact_id":{"type":"string","description":"UUID of the contact to remove (the contact's own `id`, e.g. from leadbay_research_lead_by_id's contacts list — NOT the parent lead id)."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}},{"name":"leadbay_pin_contact","annotations":{"title":"Pin a contact","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$a1","inputSchema":{"type":"object","required":["contact_id"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"contact_id":{"type":"string","description":"UUID of the contact to pin (the contact's own `id` — NOT the parent lead id)."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}},{"name":"leadbay_unpin_contact","annotations":{"title":"Unpin a contact","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$a2","inputSchema":{"type":"object","required":["contact_id"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"contact_id":{"type":"string","description":"UUID of the contact to unpin (the contact's own `id` — NOT the parent lead id)."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}},{"name":"leadbay_update_contact","annotations":{"title":"Update a contact","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":true},"description":"$a3","inputSchema":{"type":"object","required":["contact_id","first_name","last_name"],"properties":{"email":{"type":["string","null"],"description":"Contact email. Omit it to leave it unchanged. Pass null to ERASE it — an erase rewrites the whole contact, so that call must also carry every other optional field, or it is refused."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"job_title":{"type":["string","null"],"description":"Contact job title. Omit it to leave it unchanged. Pass null to ERASE it — an erase rewrites the whole contact, so that call must also carry every other optional field, or it is refused."},"last_name":{"type":"string","description":"Contact last name — REQUIRED even on an edit. Pass the current value if you're not changing it."},"contact_id":{"type":"string","description":"UUID of the contact to edit (the contact's own `id` — NOT the parent lead id)."},"first_name":{"type":"string","description":"Contact first name — REQUIRED even on an edit. Pass the current value if you're not changing it."},"phone_number":{"type":["string","null"],"description":"Contact phone (free-form). Omit it to leave it unchanged. Pass null to ERASE it — an erase rewrites the whole contact, so that call must also carry every other optional field, or it is refused."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."},"linkedin_page":{"type":["string","null"],"description":"Contact LinkedIn URL. Omit it to leave it unchanged. Pass null to ERASE it — an erase rewrites the whole contact, so that call must also carry every other optional field, or it is refused."}},"additionalProperties":false}},{"name":"leadbay_enrich_contacts","annotations":{"title":"Enrich contacts for a lead","readOnlyHint":false,"openWorldHint":true,"idempotentHint":false,"destructiveHint":true},"description":"$a4","inputSchema":{"type":"object","required":["leadId","contactId"],"properties":{"email":{"type":"boolean","description":"Enrich email address (default: true)"},"phone":{"type":"boolean","description":"Enrich phone number (default: true)"},"leadId":{"type":"string","description":"Lead UUID (required)"},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"contactId":{"type":"string","description":"Contact UUID (required)"},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}},{"name":"leadbay_create_custom_field","annotations":{"title":"Create CRM custom field","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$a5","inputSchema":{"type":"object","required":["name","_triggered_by"],"properties":{"name":{"type":"string","description":"User-visible custom field name, e.g. 'HubSpot Contact'."},"type":{"type":"string","description":"Custom field type: TEXT, NUMBER, PRICE, DATE, DATETIME, or EXTERNAL_ID. Defaults to TEXT."},"config":{"type":["object","string","null"],"description":"Type-specific config, as an object (preferred) or a JSON string. EXTERNAL_ID requires {url_template:'https://.../{value}'}; PRICE requires {currency:'USD'}; DATE/DATETIME may set {format}."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"if_not_exists":{"type":"boolean","description":"Default true. If a custom field with the same name already exists, return it instead of creating a duplicate."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["id","name","type","mapping_value","existed"],"properties":{"id":{"type":"string"},"name":{"type":"string"},"type":{"type":"string"},"config":{"type":["object","null"]},"existed":{"type":"boolean","description":"True when if_not_exists reused an existing custom field."},"mapping_value":{"type":"string","description":"Wire mapping value to use in import mappings, e.g. CUSTOM.123."}}}},{"name":"leadbay_update_custom_field","annotations":{"title":"Update CRM custom field","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":true},"description":"$a6","inputSchema":{"type":"object","required":["id","_triggered_by"],"properties":{"id":{"type":"string","description":"Custom field id to update (the numeric id from leadbay_list_mappable_fields, NOT the 'CUSTOM.<id>' mapping value)."},"name":{"type":"string","description":"New user-visible name. Omit to keep the current name."},"type":{"type":"string","description":"New type: TEXT, NUMBER, PRICE, DATE, DATETIME, or EXTERNAL_ID. Omit to keep the current type."},"config":{"type":["object","string","null"],"description":"New type-specific config, as an object (preferred) or a JSON string. EXTERNAL_ID requires {url_template:'https://.../{value}'}; PRICE requires {currency:'USD'}; DATE/DATETIME may set {format}. Omit to keep current config."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["id","name","type","mapping_value"],"properties":{"id":{"type":"string"},"name":{"type":"string"},"type":{"type":"string"},"config":{"type":["object","null"]},"mapping_value":{"type":"string","description":"Wire mapping value for import mappings, e.g. CUSTOM.123."}}}},{"name":"leadbay_delete_custom_field","annotations":{"title":"Delete CRM custom field","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":true},"description":"$a7","inputSchema":{"type":"object","required":["id","_triggered_by"],"properties":{"id":{"type":"string","description":"Custom field id to delete (the numeric id from leadbay_list_mappable_fields, NOT the 'CUSTOM.<id>' mapping value)."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"confirm":{"type":"boolean","description":"Must be true to actually delete. Without it the tool returns the field that WOULD be deleted and does nothing — a safety gate, because deletion drops the field's values from every lead."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["id","deleted"],"properties":{"id":{"type":"string"},"hint":{"type":"string","description":"Guidance — present when confirm was missing (re-call with confirm:true)."},"name":{"type":"string"},"type":{"type":"string"},"deleted":{"type":"boolean","description":"True when the field was deleted; false when confirm was not set."}}}},{"name":"leadbay_set_qualification_questions","annotations":{"title":"Modify the org's qualification questions","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":true},"description":"$a8","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"add":{"type":"array","items":{"type":"string","maxLength":255},"description":"Questions to append to the current list (deduped). Mutually exclusive with `questions`. A question Leadbay can score has ALL of: the estimative marker — it starts \"Is the company likely to \" in English or \"L'entreprise est-elle susceptible de/d' \" in French (never \"Does the company ...\" or a bare \"L'entreprise est-elle <X> ?\", which the scorer cannot verify and marks no); ONE dimension, not two joined by AND; something estimable from the company's public material, never its budget or its internal plans; and enough bite to split companies roughly 30/70 — a question nearly everyone answers yes to (\"has a website\") adds no signal, so say that and propose a sharper one rather than writing it. Max 120 chars, in the user's language. Never name a specific company in a question. Read leadbay_get_qualification_questions first: skip anything an existing question already tests, and anything the lens already filters by sector, headcount or territory."},"remove":{"type":"array","items":{"type":"string"},"description":"Exact question strings to remove from the current list. Mutually exclusive with `questions`. A removal requires confirm:true."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"confirm":{"type":"boolean","description":"Required whenever the change DROPS ANY existing question — including a same-count swap or a `questions` replacement that omits a current question, not only when the list gets shorter (removing a question changes how every lead is scored). Without it, such a change is previewed and not applied. Pure additions never need confirm."},"questions":{"type":"array","items":{"type":"string","maxLength":255},"description":"$a9"},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"add_anti_patterns":{"type":"array","items":{"type":"string"},"description":"Negative criteria to append to the org's ideal buyer profile: kinds of company that are never the buyer (\"Consulting firms and IT services companies\", \"Franchise locations of national chains\"). Qualification reads them as negative signals when it scores a lead. Uses no question slot. Its own write: do not combine with questions/add/remove. Admin only. Read leadbay_get_qualification_questions first and skip anything an existing anti-pattern or question already says. In the user's language."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["qualification_questions","count","changed"],"properties":{"hint":{"type":"string","description":"Operator note — confirm prompt on a shrink, or a no-op explanation."},"_meta":{"type":"object"},"count":{"type":"number"},"region":{"type":"string"},"changed":{"type":"boolean","description":"True when the list was actually written; false on a no-op or an unconfirmed shrink."},"anti_patterns":{"type":"array","items":{"type":"string"},"description":"The ideal buyer profile's negative criteria after an add_anti_patterns call."},"form_warnings":{"type":"array","items":{"type":"string"},"description":"Present when a written question will score poorly — not in the estimative 'Is the company likely to ...' form, or over 120 chars. The change WAS applied; tell the user and offer the rewrite."},"previous_count":{"type":"number"},"anti_patterns_added":{"type":"array","items":{"type":"string"},"description":"The criteria this call wrote."},"qualification_questions":{"type":"array","items":{"type":"object"},"description":"The questions AFTER the change. Each: {question}."}}}},{"name":"leadbay_add_note","annotations":{"title":"Add a note on a lead","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":false},"description":"Add a note to a lead. Notes are visible to the whole organization in Leadbay.\n\nWHEN TO USE: low-level — for free-form notes not tied to outreach actions, including meaningful per-lead notes/context preserved from an imported file after the import returns lead IDs. Also to record why a company was set aside when a CRM status change carries no reason, or when the user tells you something about one company that only they know (their own sales agent, already a client) without excluding it. A dislike saves its own reason: do not add a second note for it.\n\nWHEN NOT TO USE: to log an outreach action — use leadbay_report_outreach, which requires verification (gmail/calendar/user_confirmed) to prevent hallucinated outreach poisoning the SDR pipeline.\n\nThis tool MUTATES state. The caller (agent or human-in-the-loop) is responsible for confirming intent before invocation; the MCP server does not soft-prompt for confirmation. See `annotations.destructiveHint`.\n","inputSchema":{"type":"object","required":["leadId","note"],"properties":{"note":{"type":"string","description":"Note text (max 4095 characters)"},"leadId":{"type":"string","description":"Lead UUID (required)"},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["id","note","created_at"],"properties":{"id":{"type":"string","description":"Note id assigned by the backend."},"note":{"type":"string","description":"Echoed note text (truncated to 4095 chars)."},"created_at":{"type":"string","description":"ISO timestamp of creation."}}}},{"name":"leadbay_like_lead","annotations":{"title":"Like a lead","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$aa","inputSchema":{"type":"object","required":["lead_id"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"lead_id":{"type":"string","description":"UUID of the lead to like."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}},{"name":"leadbay_dislike_lead","annotations":{"title":"Dislike a lead","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$ab","inputSchema":{"type":"object","required":["lead_id"],"properties":{"reason":{"type":"string","description":"Why the lead is rejected, in the user's own words (\"société de conseil, hors ICP\", \"unpaid invoice with them\"). In a run with no user, the criterion you applied. Saved as a note on the lead that the whole team sees. Omit when no reason was given; never invent one."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"lead_id":{"type":"string","description":"UUID of the lead to dislike."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}},{"name":"leadbay_set_lead_status","annotations":{"title":"Set lead CRM status","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$ac","inputSchema":{"type":"object","required":["lead_ids","status"],"properties":{"status":{"type":"string","description":"One of: WANTED, WON, LOST, UNWANTED (case-insensitive). DEFAULT and INBOUND are accepted but are normally set by Leadbay itself."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"lead_ids":{"type":"array","items":{"type":"string"},"description":"Lead UUIDs (1-200). Every lead gets the same status."},"status_date":{"type":"string","description":"Optional YYYY-MM-DD — the date the status was actually reached (close date). Omit to let the backend stamp now."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}},{"name":"leadbay_create_campaign","annotations":{"title":"Create a named campaign (optionally seeded with leads)","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":false},"description":"$ad","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"name":{"type":"string","description":"Campaign display name (max 255 chars). Omit to let the backend AI-generate one from the seed lead_ids. If lead_ids is empty AND name is omitted, the backend assigns a default."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"lead_ids":{"type":"array","items":{"type":"string"},"description":"Lead UUIDs to attach at creation. Empty array (default) creates an empty campaign — add leads later via leadbay_add_leads_to_campaign. Non-empty seed enables AI name suggestion."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["id","name","created_at"],"properties":{"id":{"type":"string","description":"Campaign UUID."},"name":{"type":"string"},"archived":{"type":"boolean"},"created_at":{"type":"string"},"created_by":{"type":"string"},"updated_at":{"type":"string"},"ai_name_count":{"type":"number"},"last_accessed_at":{"type":"string"},"ai_generated_name":{"type":["string","null"]}}}},{"name":"leadbay_add_leads_to_campaign","annotations":{"title":"Add leads to an existing campaign","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$ae","inputSchema":{"type":"object","required":["campaign_id","lead_ids","_triggered_by"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"lead_ids":{"type":"array","items":{"type":"string"},"minItems":1,"description":"Lead UUIDs to add. Backend rejects unknown lead UUIDs with 404 — pass UUIDs sourced from pull_leads / pull_followups / tour_plan / research."},"campaign_id":{"type":"string","description":"Campaign UUID (from leadbay_create_campaign or leadbay_list_campaigns)."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["added","already_present"],"properties":{"added":{"type":"number","description":"Leads newly attached."},"already_present":{"type":"number","description":"Leads that were already in this campaign — no-op."}}}},{"name":"leadbay_remove_leads_from_campaign","annotations":{"title":"Remove leads from a campaign","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":true},"description":"$af","inputSchema":{"type":"object","required":["campaign_id","lead_ids","_triggered_by"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"lead_ids":{"type":"array","items":{"type":"string"},"minItems":1,"description":"Lead UUIDs to remove. Pass IDs sourced from Leadbay tools; invalid IDs are handled by the backend."},"campaign_id":{"type":"string","description":"Campaign UUID (from leadbay_create_campaign or leadbay_list_campaigns)."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["removed"],"properties":{"removed":{"type":"number","description":"Number of leads submitted for removal (backend returns 204, no per-lead breakdown)."}}}},{"name":"leadbay_extend_lens","annotations":{"title":"Extend a lens with additional leads (extra refill)","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":false},"description":"$b0","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"lensId":{"type":"number","description":"Lens to extend. Defaults to the user's last-active lens."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"extra_count":{"type":"number","description":"How many extra leads to request. Omit to use the backend default. The full requested count must fit into the remaining daily LENS_EXTRA_REFILL quota — otherwise the call is rejected outright (status: quota_exceeded)."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"seed_lead_ids":{"type":"array","items":{"type":"string"},"description":"Optional list of lead UUIDs from leadbay_list_lens_seed_candidates to bias the recommender. Omit or empty array → default-strategy fallback (same behaviour as a normal fill)."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["status","lens"],"properties":{"lens":{"type":"object","properties":{"id":{"type":"number"}}},"quota":{"type":"object","properties":{"resets_at":{"type":["string","null"]},"used_today":{"type":["number","null"]}},"description":"Only present on status=quota_exceeded. Shows the org's daily LENS_EXTRA_REFILL state."},"reason":{"type":"object","required":["code","message","retryable"],"properties":{"code":{"type":"string","description":"audience_too_narrow (lens holds nothing and its criteria intersect to nothing) | no_new_leads (everything matching has already been delivered) | no_candidates (pool is empty; which of the two could not be determined)"},"message":{"type":"string"},"criteria":{"type":"object","description":"The criteria in play, so the agent can name them."},"retryable":{"type":"boolean","description":"Always false here. Re-calling leadbay_extend_lens cannot change the outcome — widen the audience instead."},"narrow_locations":{"type":"array","items":{"type":"object"},"description":"Include-locations that resolved to city-scale or smaller — the usual thing to relax first."}},"description":"Only present on status=no_candidates. Same shape and `code` vocabulary as leadbay_pull_leads' empty_reason, so one routing rule covers both tools."},"status":{"type":"string","description":"queued | no_candidates | quota_exceeded | refresh_in_progress | no_valid_seeds"},"message":{"type":"string","description":"Human-readable summary. On error statuses, this is the line to surface to the user."},"accepted_seeds":{"type":"array","items":{"type":"string"},"description":"Subset of seed_lead_ids that passed validation and will bias the fill. Empty when no seeds were submitted (default-strategies fallback). Present only on status=queued."},"available_count":{"type":["number","null"],"description":"How many leads a refill on this lens could still draw, read from /extra_refill_preview before queueing. 0 means the refill was NOT queued (status=no_candidates). null means the pool could not be read and the refill was queued anyway."}}}},{"name":"leadbay_manage_lenses","annotations":{"title":"List, switch, edit, or delete your lenses","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":true},"description":"$b1","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"confirm":{"type":"boolean","description":"Required (=true) to actually delete. Without it, deleteLensId returns a preview to confirm with the user first."},"newName":{"type":"string","description":"New lens name (used with editLensId)."},"editLensId":{"type":["string","number"],"description":"When set, edit this lens's metadata — provide newName and/or newDescription. Must be one of the user's lenses."},"deleteLensId":{"type":["string","number"],"description":"When set, delete this lens. DESTRUCTIVE — returns a delete_preview unless confirm:true. Cannot delete the default lens."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"newDescription":{"type":"string","description":"New lens description (used with editLensId). Pass an empty string to clear it."},"switchToLensId":{"type":["string","number"],"description":"When set, switch the active lens to this id (must be one of the user's lenses), then return the refreshed list."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["status","lenses","active_lens_id"],"properties":{"edited":{"type":"boolean","description":"True when this call renamed/re-described a lens."},"lenses":{"type":"array","items":{"type":"object"},"description":"The user's lenses, each with every field GET /lenses returns plus is_active, is_default and `criteria` (its sectors, locations and sizes by name; null if unreadable)."},"status":{"type":"string","description":"'listed', 'switched', 'already_active', 'edited', 'deleted', 'delete_preview' (confirm to proceed), 'cannot_delete_default', or 'not_found'."},"deleted":{"type":"boolean","description":"True when this call deleted a lens."},"message":{"type":"string"},"switched":{"type":"boolean","description":"True when this call changed the active lens."},"will_delete":{"type":"object","description":"On 'delete_preview': the lens that WILL be deleted {id, name}. Nothing removed yet."},"active_lens_id":{"type":["string","null"]}}}},{"name":"leadbay_new_lens","annotations":{"title":"Create a new named lens","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":false},"description":"$b2","inputSchema":{"type":"object","required":["name","_triggered_by"],"properties":{"base":{"type":"number","description":"Lens id to clone from. Defaults to the active/default lens."},"name":{"type":"string","maxLength":255,"description":"Display name for the new lens (required)."},"sizes":{"type":"array","items":{"type":"object","properties":{"max":{"type":"number"},"min":{"type":"number"}}},"description":"Company size buckets, e.g. [{min:30,max:300}]."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"confirm":{"type":"boolean","description":"Safety gate. Defaults to false → the tool returns a PREVIEW and creates nothing. Show the preview to the user, get their explicit go-ahead, then re-call the SAME args with confirm:true to actually create the lens."},"sectors":{"type":"array","items":{"type":"string"},"description":"Sectors to include — free text (auto-resolved) or ids."},"locations":{"type":"array","items":{"type":"string"},"description":"Geographic scope — free text (e.g. ['Indre-et-Loire', 'Texas']) or admin-area ids. Resolved via /geo/search at any level from state down to city (state / région / département / county / city). NEVER a country name — this workspace serves exactly ONE country, so a whole-country ask means passing NO location at all (rejected with COUNTRY_LEVEL_LOCATION). Scopes the lens to a sales territory."},"description":{"type":"string","maxLength":255,"description":"Optional lens description."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"exclude_sectors":{"type":"array","items":{"type":"string"},"description":"Sectors to exclude — free text or ids."},"exclude_locations":{"type":"array","items":{"type":"string"},"description":"Locations to exclude — free text or ids. Sub-country areas only — excluding a country is meaningless on a single-country workspace and is rejected."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["status"],"properties":{"lens":{"type":"object","description":"On 'created': the created lens {id, name}."},"_meta":{"type":"object"},"status":{"type":"string","description":"'preview', 'created', 'ambiguous_sectors', 'ambiguous_locations', 'country_level_location', or 'orphan_created' (filter write failed + cleanup failed)."},"message":{"type":"string"},"will_create":{"type":"object","description":"On 'preview': what WILL be created — {name, description, sectors, exclude_sectors, sizes, locations, exclude_locations}. Nothing has been written yet."},"filter_applied":{"type":"object","description":"On 'created': the FilterPayload POSTed to the new lens."},"filter_preview":{"type":"object","description":"On 'preview': the FilterPayload that would be applied."},"country_locations":{"type":"array","items":{"type":"object"},"description":"On 'country_level_location': per offending value {value, param, kind, country, axis, kept}. A country name is never a location criterion — each workspace serves exactly ONE country. The recovery BRANCHES on `country_locations[].axis` and `[].kind`; `hint` states the one for THIS call — follow it verbatim. When the country was the ONLY scope, or on ANY non-foreign `exclude`, the answer is to write NOTHING at all — re-calling with the value merely dropped persists a scope that inverts the request. Never retry with another spelling or a nearby city."},"computing_wishlist":{"type":"boolean","description":"On 'created': true when the lens has criteria and its wishlist is (re)computing asynchronously — an immediate leadbay_pull_leads may read empty; wait ~30s. False for a criteria-less clone (it inherits the base lens's leads immediately)."},"sector_ambiguities":{"type":"array","items":{"type":"object"},"description":"On 'ambiguous_sectors': per text {sector_text, matches:[{id,name,score}]}."},"location_ambiguities":{"type":"array","items":{"type":"object"},"description":"On 'ambiguous_locations': per text {location_text, matches:[{id,name,country,level,score}]}. Re-call the chosen id via the SAME axis the text came from — an include text → locations; a text from exclude_locations → exclude_locations (NOT locations, which would include the area the user asked to exclude). The `message` field names the correct param per text."}},"description":"'preview' (default, NOTHING created — confirm with the user then re-call with confirm:true); 'created' on success; 'ambiguous_sectors' / 'ambiguous_locations' when free-text sectors / locations didn't resolve (re-call with ids — the lens was NOT created); 'country_level_location' when a country-level value was passed as a location (the lens was NOT created; read `hint` — re-calling without the value is often itself wrong)."}},{"name":"leadbay_set_telemetry","annotations":{"title":"Enable, disable, or check product-usage telemetry","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$b3","inputSchema":{"type":"object","properties":{"action":{"enum":["enable","disable","status"],"type":"string","description":"enable / disable flip telemetry for the user; status just reports the current setting. Defaults to status."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}}],"prompts":[{"name":"leadbay_daily_check_in","arguments":[],"description":"Morning DISCOVERY workflow — new leads from the lens wishlist. Trigger\non \"show me leads\", \"what's new today\", \"let's prospect\", \"run my check-in\",\n\"my morning check-in\", \"I do this every day\", \"every morning\". Recurrence\nlanguage always means this prompt. Do NOT trigger on follow-up phrasings\n(\"follow up\", \"before my trip\") — those go to `leadbay_followup_check_in`.\n"},{"name":"leadbay_prospecting_overview","arguments":[],"description":"Orientation for working with Leadbay from any host — discovery vs.\nfollow-up, the outreach loop, outcome recording, imports, pushback /\nsnooze, and the connected-outreach-tool registry. Trigger when the\nconversation involves Leadbay leads, prospecting, pipeline, follow-up,\noutreach, or lens / ICP — anything from \"show me my leads\" to \"what\nshould I follow up on\" to \"I'll send via lemlist\".\n"},{"name":"leadbay_new_leads","arguments":[{"name":"need","required":false,"description":"What the user is looking for, in their own words (e.g. '10 gyms around Dallas that would buy modular flooring, with phone numbers'). Optional — the session starts by asking when absent."}],"description":"Guided net-new lead delivery — turn a described need (\"gyms around Dallas\nthat would buy our flooring\") into ICP-perfect NEW companies with\nqualification evidence and the right contact, via leadbay_find_new_leads.\nTrigger when the user DESCRIBES who they want: \"get me N companies that\n<profile>\", \"we're entering <market>\". A bare \"find me new leads\" with no\nprofile, and \"today's leads\", are the daily lens batch — leadbay_pull_leads.\n\"Qualify these companies I have\" is leadbay_qualify_leads.\n"},{"name":"leadbay_research_a_domain","arguments":[{"name":"domain","required":true,"description":"Company name or domain (for example 'Acme Corporation' or 'acme.com'). The legacy argument key remains `domain` for client compatibility."}],"description":"Resolve a company by name or domain across the user's visible Discover,\nMonitor, and Activate corpus, then return everything Leadbay knows about it.\n"},{"name":"leadbay_import_file","arguments":[{"name":"file","required":false,"description":"Path or user-visible name of the CSV/file to import. If omitted, use the file the user attached or referenced."},{"name":"instruction","required":false,"description":"Additional user goal, e.g. \"then qualify the leads\", \"preserve owner phone as a custom field\", or \"only import restaurants in Manhattan\"."}],"description":"Import a user-supplied CSV/file into Leadbay through five phases with\nevidence gates — scan, derive, resolve identities, preserve & commit,\nthen optionally qualify and report. The job is to maximize how many\nrows the Leadbay system actually ingests and matches.\n"},{"name":"leadbay_refine_audience","arguments":[{"name":"instruction","required":true,"description":"The refinement (e.g. 'focus on hospitals running their own IT'). Set to plain English."}],"description":"Refine the kind of leads Leadbay surfaces beyond firmographics, with a\nfree-text instruction. Handles the clarification round-trip if the new\nprompt is ambiguous.\n"},{"name":"leadbay_log_outreach","arguments":[{"name":"lead_id","required":true,"description":"The lead UUID. Get it from leadbay_pull_leads or leadbay_research_lead_by_id."},{"name":"summary","required":true,"description":"1-2 sentences describing what I did (e.g. 'Sent intro email to CTO citing recent Hornsea contract')."}],"description":"Log outreach (an email I sent, a call I made, a meeting I had) on a\nspecific lead. Captures verification so the SDR pipeline trusts the entry.\n"},{"name":"leadbay_sync_outreach","arguments":[{"name":"lookback_days","required":false,"description":"Optional: how many days back the first run reads (default 14). The daily task always reads the last 2 days."}],"description":"Keep Leadbay's record of who was contacted current: read my mailbox and\ncalendar, log every email and meeting with a Leadbay lead on the person it\ninvolved, then schedule the same sync to run every day. Trigger on \"log my\nemails in Leadbay\", \"sync my outreach\", \"keep Leadbay up to date with my\nemails\".\n"},{"name":"leadbay_plan_tour_in_city","arguments":[{"name":"city","required":true,"description":"City or region the user is visiting (e.g. 'Limoges', 'Bay Area'). Used as the geo filter for both Monitor and Discover lookups. A country is not a city: this workspace already covers exactly one country, and a country name here silently fences the tour to a same-named village. Do NOT omit the argument to recover — a city-less tour returns arbitrary leads from across the whole workspace, which is not an itinerary. Ask which city or region the visit is to."},{"name":"date","required":false,"description":"When the visit is (e.g. 'May 24', 'next Thursday'). Surfaced in the outreach drafts as 'I'll be in <city> on <date>'."}],"description":"Use whenever the user names a city they'll be in and asks who to see\n— \"I'm in SF next Tuesday, who's worth meeting?\", \"I'm going to Berlin\n— who should I visit?\", \"plan my <city> tour\". Any in-person/visit\nintent tied to a place routes here, NOT to `leadbay_pull_leads`. It\nsurfaces follow-ups + fresh Discover leads in the city via\n`leadbay_tour_plan`, ALWAYS offers to plot them on a map (rendering it\non yes), then offers outreach drafts + campaign persistence.\n"},{"name":"leadbay_build_campaign","arguments":[{"name":"audience","required":false,"description":"Optional: a fresh audience to target (e.g. 'dental clinics in Texas'). Omit to build from your ACTIVE lens — the default."},{"name":"campaign_name","required":false,"description":"Optional: a name for the campaign. Omit and one is derived from the lens/audience + date (or the backend AI-names it)."},{"name":"count","required":false,"description":"Optional: how many fully-actionable leads to build (default 20). The loop keeps discovering, qualifying and enriching until this many in-ICP leads each have a reachable target-title contact — or the lens is exhausted. Higher counts take longer and consume more quota."},{"name":"job_titles","required":false,"description":"Optional: the exact buyer job titles to enrich, comma-separated (e.g. 'VP Sales, Head of Growth, Director of Business Development'). Omit and the buyer persona is derived from what you sell. A lead only counts toward the target when it has a reachable contact matching one of these titles."}],"description":"Build a sales campaign from scratch, autonomously, to a target size:\ndiscover on the lens, qualify, and enrich the buyer titles until `count`\nleads each have a reachable target-title contact — no pauses, no confirm\ngates. Saves via `leadbay_create_campaign` and renders a one-tap\ncall/email view via `leadbay_campaign_call_sheet`. Trigger on \"build me a\ncampaign\", \"build N leads\", \"create a campaign from scratch\". Work an\nexisting one with `leadbay_work_campaign`.\n"},{"name":"leadbay_setup_team_prospecting","arguments":[{"name":"audience","required":true,"description":"Natural-language audience description (e.g. 'plumbing companies with 10-50 employees in Seine-Maritime'). The lens-creation step (`leadbay_refine_lead_targeting` → `leadbay_create_lens`) interprets it. A country name is not a scope here — this workspace already covers exactly one country, so drop it and keep the rest of the description; a DIFFERENT country cannot be targeted at all."},{"name":"rep_split","required":false,"description":"Optional: how to split the validated leads into per-rep campaigns. Free text — e.g. 'split by city' or 'one campaign per rep: John gets Tulsa, Sarah gets OKC'. Splitting by country is not a split — the workspace is single-country."}],"description":"Manager-led prospecting setup: conversationally turn a natural-language\naudience ask into a Leadbay lens, validate the candidate leads, and\npersist them as one or more named campaigns the rep(s) can work\nthrough. Closes #3630 US3 end-to-end (within the current\ncreator-scoped campaign visibility model).\n"},{"name":"leadbay_work_campaign","arguments":[{"name":"campaign","required":false,"description":"Campaign name (fuzzy match against your own campaigns) or campaign UUID. Omit to list and pick interactively."},{"name":"mode","required":false,"description":"Optional: skip the readiness-assessment proposal and jump directly into 'call_sheet' / 'email_sheet' / 'map' / 'enrich_first'. Omit (recommended) and let the prompt propose based on the data."}],"description":"Work a campaign as a real outreach session: pick the campaign,\nassess what the user has (phones / emails / coords), then PROPOSE\nthe right session mode (call sheet, email sheet, enrich titles\nfirst, map). After they pick, render — and as they dictate\noutcomes per lead, record both note + epilogue via\n`leadbay_report_outreach` in one round trip.\n"},{"name":"leadbay_outreach_session","arguments":[{"name":"source","required":false,"description":"Optional: where the leads come from — 'followups' (the Monitor, default), 'discover' (the active lens), or 'campaign'. Omit and the session asks."},{"name":"campaign","required":false,"description":"Optional: campaign name (fuzzy-matched against your own campaigns) or UUID, when source is 'campaign'. Omit to list and pick."}],"description":"Run an outreach session over ANY set of leads: follow-ups, a Discover\nlens, or a campaign. Assess who is actually reachable, build the lead\ndesk, then record each outcome as you dictate it. Trigger on \"relance\nplan\", \"campagne de relance\", \"cold calling session\", \"prospection\nsession\", \"outreach session\", \"let's contact leads\", \"who do I call\ntoday\". For a campaign the user NAMES, prefer `leadbay_work_campaign`.\n"},{"name":"leadbay_route_session","arguments":[{"name":"city","required":false,"description":"Where the rep is going — a city or region (e.g. 'Lyon', 'Hauts-de-France'). Omit and the session asks. NEVER a country: this workspace serves one country, so a country name is not a place to tour."},{"name":"day","required":false,"description":"Optional: which day the trip is, in the rep's own words ('Thursday', 'next week'). Used only to frame the plan; nothing is scheduled."}],"description":"Plan and work a day on the road: pick the place, see the leads on a map,\ndrive them in a sensible order and record each visit as it happens.\nTrigger on \"plan my route\", \"tournée\", \"road trip\", \"I'm driving to\n<city>\", \"who can I visit in <city>\", \"plan my day in the field\",\n\"prospection terrain\". For working leads by PHONE from a desk, prefer\n`leadbay_outreach_session`.\n"},{"name":"leadbay_qualify_top_n","arguments":[{"name":"count","required":false,"description":"How many leads to qualify (default 10, max 25). Higher counts may take 5+ minutes."}],"description":"Bulk-qualify the top N un-qualified leads in the active lens. Uses\nleadbay_bulk_qualify_leads with a sensible default budget.\n"},{"name":"leadbay_top_accounts_to_activate","arguments":[{"name":"count","required":false,"description":"Optional: how many accounts the plan should hold (default 50)."},{"name":"territory","required":false,"description":"Optional: restrict the plan to a territory (e.g. 'Indre-et-Loire', 'Région Ouest'). Sets geography on the Discover lens. A country is not a territory — this workspace already covers exactly one country."}],"description":"Build a ranked account-conquest plan from Leadbay data — the accounts worth\nactivating, each with a motif, a pitch and a checklist, ranked by the\nstrongest Leadbay signal. Every figure carries its source, and anything\nLeadbay can't measure is shown as OMITTED rather than estimated. Uses\n`leadbay_bulk_qualify_leads` and `leadbay_enrich_titles`. Trigger on\n\"top 50 accounts to activate\", \"who should we go after\".\n"},{"name":"leadbay_getting_started","arguments":[],"description":"Guided first-run walkthrough — four clicks that actually use Leadbay: check\nthe account, pull today's leads, draft a first email to the top one, then\nreveal who to send it to. Use when the user is new or asks to be SHOWN how\nLeadbay works (\"walk me through Leadbay\", \"I'm new\", \"how do I use this\",\n\"give me a tour\"). Don't use it for orientation prose with no clicking —\nthat's leadbay_prospecting_overview.\n"}],"resources":[{"uri":"org://taste-profile","name":"Org taste profile","mimeType":"application/json","description":"The org's qualification questions, intent tags, and ICP signals — the agent's knowledge base for what makes a lead a fit."}],"triggers":[],"configSchema":null,"configAllowsEmpty":false},"className":"w-12 h-12 shrink-0"}],"$Lb4"]}],"$Lb5"]}],"$Lb6"]}]}],"$Lb7"]}],"$Lb8"]}]] </article></body></html>

61:[["$","$L6f",null,{"namespace":"leadbay","slug":"leadbay","ownerId":"org_01M3WXCFQTZKXSKCHM6Z8AY5VK","unlisted":false,"isLegacyFlyDeployment":false}],["$","$Lb",null,{"className":"mt-0 pt-4","children":[["$","div",null,{"className":"space-y-4","children":[["$","div",null,{"className":"pt-8","children":["$","div",null,{"className":"mb-3 shrink-0","children":[["$","div",null,{"className":"flex flex-col gap-3 sm:flex-row sm:items-start sm:justify-between sm:gap-4","children":[["$","div",null,{"className":"flex items-start gap-3 min-w-0 flex-1","children":[["$","$L70",null,{"server":{"id":"2271d337-d3bf-45aa-8b44-b8628e689da5","displayName":"Leadbay","description":"AI lead discovery, qualification, and outreach prep on your Leadbay account.\n\nLeadbay finds lookalikes of your best customers and reasons like your best sales rep, at scale. Connect it to Claude, ChatGPT, Cursor or any MCP client to pull leads, qualify them, draft outreach and log activity in plain language, on your real Leadbay data and with your permissions.\n\nSign in once with your Leadbay account (OAuth), no API token needed. US accounts use https://mcp.leadbay.app/mcp, France / EU accounts use https://mcp.leadbay.app/fr/mcp.\n\nDocs: https://docs.leadbay.app/doc/leadbay-mcp/what-is-leadbay-mcp","namespace":"leadbay","slug":"leadbay","qualifiedName":"leadbay/leadbay","iconUrl":"https://www.google.com/s2/favicons?domain=leadbay.ai&sz=64","homepage":"https://leadbay.ai","repositoryUrl":"https://github.com/leadbay/mcp","backlinkUrl":null,"license":null,"verified":false,"official":false,"verification":{"official":false,"homepageHost":"leadbay.ai","verificationValue":"smithery-verification=2f8ffb78741d80fbe38edd2d6e992360d64b6e03e4d12484656ed73b65599d80","hasDomainProof":false,"hasBacklinkProof":true,"backlinkUrl":"https://leadbay.ai","checkedAt":"$D2026-10-02T00:19:33.296Z","nextCheckAt":"$D2026-10-02T01:19:33.296Z"},"unlisted":false,"remote":true,"createdAt":"$D2026-10-01T23:41:55.674Z","organizationId":"org_01M3WXCFQTZKXSKCHM6Z8AY5VK","bySmithery":false,"runToolsSlug":"leadbay--leadbay","securityScan":{"isSecure":null},"lastDeployedAt":"2026-10-01T23:55:43.450Z","lastSuccessfulDeployment":{"type":"external_shttp","mcpUrl":"https://leadbay--leadbay.run.tools","upstreamUrl":"https://mcp.leadbay.app/mcp","configSchema":{},"cfDeploymentHash":null},"mcpUrl":"https://leadbay--leadbay.run.tools","tools":[{"name":"leadbay_lead_job_status","annotations":{"title":"Poll a lead-delivery job","readOnlyHint":true,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$71","inputSchema":{"type":"object","required":["job_id","_triggered_by"],"properties":{"limit":{"type":"number","description":"Items per page, 1-100 (default 100; pages are auto-collected)."},"since":{"type":"string","description":"Opaque cursor from a previous poll's next_since — returns only items emitted after it. Omit for the full snapshot."},"job_id":{"type":"string","description":"The job_id returned by leadbay_find_new_leads or leadbay_qualify_leads."},"offset":{"type":"number","description":"With compact: the first row to return (default 0). Each result carries at most 100 rows; take the next value from next_poll.offset."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"compact":{"type":"boolean","description":"true = one row per company (name, website, LinkedIn) in the order of the user's list, plus coverage counts, instead of full leads. Pass it when next_poll carries it (a leadbay_qualify_leads identity pass). Reads the whole job; since is ignored."},"wait_seconds":{"type":"number","description":"0 (default) = instant snapshot. >0 = keep polling up to this many seconds (maximum 45) until the job is terminal — use 45 when the user asked to wait for results."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false}},{"name":"leadbay_pull_leads","annotations":{"title":"Pull fresh Leadbay leads","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$72","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"page":{"type":"number","description":"Page number, 0-indexed (default 0)"},"count":{"type":"number","description":"Leads per page, max 50 (default 20)"},"order":{"type":"string","description":"Optional sort, FIELD:ASC|DESC (SCORE, NAME, SIZE, SECTOR, STATUS, CONTACT_COUNT, LAST_PROSPECTING_ACTION_AT, LIKED). Omit for the lens's own Discover ranking. An unknown value is rejected and the error lists every accepted order."},"lensId":{"type":"number","description":"Override the auto-resolved last-active lens (escape hatch — normally omit)"},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"verbose":{"type":"boolean","description":"If true, include the full set of lead-summary fields. Default false: returns the trimmed agent-friendly form."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["lens","leads","pagination"],"properties":{"lens":{"type":"object","properties":{"id":{"type":"number"}},"description":"Lens metadata (id of the lens that was queried)."},"_meta":{"type":"object","properties":{"region":{"type":"string"},"latency_ms":{"type":["number","null"]}},"description":"Operator context: region + last-call latency."},"leads":{"type":"array","items":{"type":"object"},"description":"The page of leads. In default mode (verbose:false) each lead is the trimmed agent-friendly shape; in verbose:true the full LeadPayload."},"has_more":{"type":"boolean","description":"True if at least one more page exists. Spec-aligned pagination metadata."},"next_page":{"type":["number","null"],"description":"0-indexed next page number, or null on the last page."},"next_steps":{"type":["object","null"],"properties":{"options":{"type":"array","items":{"type":"object","properties":{"kind":{"type":"string"},"label":{"type":"string"},"description":{"type":"string"}}}},"question":{"type":"string"}},"description":"$73"},"pagination":{"type":"object","properties":{"page":{"type":"number"},"pages":{"type":"number"},"total":{"type":"number"}},"description":"page (0-indexed), pages (total), total (item count)."},"empty_reason":{"type":["object","null"],"required":["code","message","retryable"],"properties":{"code":{"type":"string","description":"computing | no_candidates | no_new_leads | audience_too_narrow | unknown"},"message":{"type":"string","description":"The line to surface to the user."},"criteria":{"type":"object","description":"The lens criteria in play — what the user would have to relax. Present when the lens carries any."},"retryable":{"type":"boolean","description":"True only while the lens is still computing. False means no amount of re-pulling or extending can produce leads."},"narrow_locations":{"type":"array","items":{"type":"object"},"description":"Include-locations that resolved to a city-scale area or smaller ({id, name, level}). On an empty lens, name these first: this is the fingerprint of a whole-country location that fell through to a same-named village (product#3951)."}},"description":"Why this LENS holds zero leads. null whenever leads were returned, and null when this page is empty only because it is past the end of a non-empty lens. `retryable` is the field to route on: true ONLY on code=computing (pull again in ~30s). On every other code re-pulling and leadbay_extend_lens are both futile — a refill on a zero-candidate lens answers 'queued', consumes no quota and delivers nothing — so surface `message` to the user and offer leadbay_adjust_audience instead of retrying."},"computing_scores":{"type":"boolean","description":"True if scoring is still running."},"computing_wishlist":{"type":"boolean","description":"True if Leadbay is still rebuilding this lens's wishlist."}}}},{"name":"leadbay_pull_followups","annotations":{"title":"Pull known leads to follow up on (Monitor view)","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$74","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"city":{"type":"string","description":"Free-text city / region (e.g. 'Berlin', 'NYC', 'São Paulo'). The composite resolves it to an admin_area id via GET /geo/search and merges it into the active Monitor filter as a `location_ids` FilterCriterion. Ambiguous matches surface as `status: 'ambiguous_locations'` with `location_ambiguities[]` — the agent picks an id and re-calls via `city_id`. NEVER a country name: this workspace serves exactly ONE country, so a whole-country ask means omitting `city` entirely."},"page":{"type":"number","description":"Page number, 0-indexed (default 0)."},"count":{"type":"number","description":"Leads per page, max 200 (default 20)."},"liked":{"type":"boolean","description":"When true, restrict to leads the user has explicitly liked. Default false."},"order":{"type":"string","description":"Optional sort, FIELD:ASC|DESC (SCORE, NAME, SIZE, SECTOR, STATUS, CONTACT_COUNT, LAST_PROSPECTING_ACTION_AT, LIKED). Omit for the Monitor's own ranking. An unknown value is rejected and the error lists every accepted order."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"city_id":{"type":"string","description":"Pre-resolved admin_area id (numeric string). Use when the user / agent has already picked one of the ambiguity candidates. Bypasses the resolver."},"filtered":{"type":"boolean","description":"Apply the user's stored Monitor filter (server-persisted via POST /monitor/filter). Default true."},"personal":{"type":"boolean","description":"When true, restrict to leads this user has personally monitored (not org-wide). Default false."},"set_filter":{"type":"object","properties":{"criteria":{"type":"array","items":{"type":"object"},"description":"Array of FilterCriterion objects per the backend FilterCriterion anyOf schema (location_ids, sector_ids, size, keywords, last_action, last_action_date, liked, yc, custom_field, custom_field_comparison). A `location_ids` criterion must carry sub-country admin areas only — a country name here is rejected with COUNTRY_LEVEL_LOCATION before anything is persisted."}},"description":"Optional FilterItem ({criteria: FilterCriterion[]}). When provided, the composite POSTs it to /monitor/filter (server-persists across sessions) BEFORE fetching the filtered Monitor view. Use to refine 'leads to follow up' by city, sector, recency, action type, etc."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["leads"],"properties":{"_meta":{"type":"object","properties":{"region":{"type":"string"},"latency_ms":{"type":["number","null"]}},"description":"Operator context: region + last-call latency."},"leads":{"type":"array","items":{"type":"object"},"description":"The page of monitored leads. Each lead carries the FullLead shape augmented with normalized linkedin_page on contacts and `recommended_contact`. A lead whose latest email was logged with a Gmail message id also carries `last_logged_email: {gmail_message_id, logged_at}`."},"status":{"type":"string","description":"`ambiguous_locations` when a passed `city` matched multiple admin_areas; the agent picks an id from `location_ambiguities` and re-calls with `city_id`. `country_level_location` when `city`, `city_id` or a `set_filter` `location_ids` criterion carried a country-level value — nothing was read and no filter was persisted; read `hint` for the recovery, which differs per case. Absent on the happy path."},"next_steps":{"type":["object","null"],"properties":{"options":{"type":"array","items":{"type":"object","required":["label","description","kind"],"properties":{"kind":{"type":"string"},"label":{"type":"string"},"description":{"type":"string"}}}},"question":{"type":"string"}},"description":"Deterministic follow-on offers, artifact option FIRST. Map `options[]` into the host's next-step widget VERBATIM and in order — do not reword, reorder or drop them. null when the page is empty. Each option: {label (≤5 words), description (the full sentence), kind}."},"pagination":{"type":["object","null"],"description":"page / pages / total — the backend's pagination envelope when present."},"reply_check":{"type":"string","description":"Present when at least one lead carries `last_logged_email`: how to look for replies in the user's mailbox and log them before rendering."},"outreach_sync":{"type":"string","description":"Present when more than half the leads on the page have no outreach logged: the offer of the daily mailbox and calendar sync, and how to run it."},"active_filters":{"type":["object","null"],"description":"The FilterItem currently stored server-side for this user (via GET /monitor/filter). null when no filter is set or when filtered:false was passed."},"filter_applied":{"type":["boolean","null"],"description":"Whether the `set_filter` just sent actually landed. null when no set_filter was passed. FALSE means the backend answered 200 but stored nothing — every count and row below belongs to the PREVIOUS filter, echoed in `active_filters`, NOT to what was requested. Do not report those figures as the requested segment; re-send with a `type` discriminator on each criterion."},"country_locations":{"type":"array","items":{"type":"object"},"description":"Per offending value: {value, param, kind, country, axis, kept}. Only present when `status === 'country_level_location'`. The recovery BRANCHES on `country_locations[].axis` and `[].kind`; `hint` states the one for THIS call — follow it verbatim. Dropping the argument is NOT the general answer: on an `exclude` axis it returns the very companies the user asked to remove, and for a `foreign_country` an unfiltered result is this workspace's own leads, which answer a different question. Never retry with another spelling or a nearby city."},"location_ambiguities":{"type":"array","items":{"type":"object"},"description":"Per ambiguous city: {location_text, matches:[{id, name, country, level, score}]}. Only present when `status === 'ambiguous_locations'`."},"total_excluded_by_pushback":{"type":"number","description":"Composite-derived count of leads in the page that were excluded because their `pushback_status` is active. The backend may or may not pre-filter; this exposes the count when the composite has to drop them itself."}}}},{"name":"leadbay_followups_map","annotations":{"title":"Plot follow-up leads on a map (travel / in-person intent)","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$75","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"city":{"type":"string","description":"Free-text city / region (e.g. 'Berlin', 'NYC', 'São Paulo'). The composite resolves it to an admin_area id via GET /geo/search and merges it into the active Monitor filter as a `location_ids` FilterCriterion. Ambiguous matches surface as `status: 'ambiguous_locations'` with `location_ambiguities[]` — the agent picks an id and re-calls via `city_id`. NEVER a country name: this workspace serves exactly ONE country, so a whole-country ask means omitting `city` entirely."},"page":{"type":"number","description":"Page number, 0-indexed (default 0)."},"count":{"type":"number","description":"Leads per page, max 200 (default 20)."},"liked":{"type":"boolean","description":"When true, restrict to leads the user has explicitly liked. Default false."},"order":{"type":"string","description":"Optional sort, FIELD:ASC|DESC (SCORE, NAME, SIZE, SECTOR, STATUS, CONTACT_COUNT, LAST_PROSPECTING_ACTION_AT, LIKED). Omit for the Monitor's own ranking. An unknown value is rejected and the error lists every accepted order."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"city_id":{"type":"string","description":"Pre-resolved admin_area id (numeric string). Use when the user / agent has already picked one of the ambiguity candidates. Bypasses the resolver."},"filtered":{"type":"boolean","description":"Apply the user's stored Monitor filter (server-persisted via POST /monitor/filter). Default true."},"personal":{"type":"boolean","description":"When true, restrict to leads this user has personally monitored (not org-wide). Default false."},"set_filter":{"type":"object","properties":{"criteria":{"type":"array","items":{"type":"object"},"description":"Array of FilterCriterion objects per the backend FilterCriterion anyOf schema (location_ids, sector_ids, size, keywords, last_action, last_action_date, liked, yc, custom_field, custom_field_comparison). A `location_ids` criterion must carry sub-country admin areas only — a country name here is rejected with COUNTRY_LEVEL_LOCATION before anything is persisted."}},"description":"Optional FilterItem ({criteria: FilterCriterion[]}). When provided, the composite POSTs it to /monitor/filter (server-persists across sessions) BEFORE fetching the filtered Monitor view. Use to refine 'leads to follow up' by city, sector, recency, action type, etc."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["leads"],"properties":{"_meta":{"type":"object","properties":{"region":{"type":"string"},"latency_ms":{"type":["number","null"]}},"description":"Operator context: region + last-call latency."},"leads":{"type":"array","items":{"type":"object"},"description":"The page of monitored leads. Each lead carries the FullLead shape augmented with normalized linkedin_page on contacts and `recommended_contact`. A lead whose latest email was logged with a Gmail message id also carries `last_logged_email: {gmail_message_id, logged_at}`."},"status":{"type":"string","description":"`ambiguous_locations` when a passed `city` matched multiple admin_areas; the agent picks an id from `location_ambiguities` and re-calls with `city_id`. `country_level_location` when `city`, `city_id` or a `set_filter` `location_ids` criterion carried a country-level value — nothing was read and no filter was persisted; read `hint` for the recovery, which differs per case. Absent on the happy path."},"next_steps":{"type":["object","null"],"properties":{"options":{"type":"array","items":{"type":"object","required":["label","description","kind"],"properties":{"kind":{"type":"string"},"label":{"type":"string"},"description":{"type":"string"}}}},"question":{"type":"string"}},"description":"Deterministic follow-on offers, artifact option FIRST. Map `options[]` into the host's next-step widget VERBATIM and in order — do not reword, reorder or drop them. null when the page is empty. Each option: {label (≤5 words), description (the full sentence), kind}."},"pagination":{"type":["object","null"],"description":"page / pages / total — the backend's pagination envelope when present."},"reply_check":{"type":"string","description":"Present when at least one lead carries `last_logged_email`: how to look for replies in the user's mailbox and log them before rendering."},"outreach_sync":{"type":"string","description":"Present when more than half the leads on the page have no outreach logged: the offer of the daily mailbox and calendar sync, and how to run it."},"active_filters":{"type":["object","null"],"description":"The FilterItem currently stored server-side for this user (via GET /monitor/filter). null when no filter is set or when filtered:false was passed."},"filter_applied":{"type":["boolean","null"],"description":"Whether the `set_filter` just sent actually landed. null when no set_filter was passed. FALSE means the backend answered 200 but stored nothing — every count and row below belongs to the PREVIOUS filter, echoed in `active_filters`, NOT to what was requested. Do not report those figures as the requested segment; re-send with a `type` discriminator on each criterion."},"country_locations":{"type":"array","items":{"type":"object"},"description":"Per offending value: {value, param, kind, country, axis, kept}. Only present when `status === 'country_level_location'`. The recovery BRANCHES on `country_locations[].axis` and `[].kind`; `hint` states the one for THIS call — follow it verbatim. Dropping the argument is NOT the general answer: on an `exclude` axis it returns the very companies the user asked to remove, and for a `foreign_country` an unfiltered result is this workspace's own leads, which answer a different question. Never retry with another spelling or a nearby city."},"location_ambiguities":{"type":"array","items":{"type":"object"},"description":"Per ambiguous city: {location_text, matches:[{id, name, country, level, score}]}. Only present when `status === 'ambiguous_locations'`."},"total_excluded_by_pushback":{"type":"number","description":"Composite-derived count of leads in the page that were excluded because their `pushback_status` is active. The backend may or may not pre-filter; this exposes the count when the composite has to drop them itself."}}}},{"name":"leadbay_tour_plan","annotations":{"title":"Plan a mixed-mode tour itinerary (known + fresh leads)","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$76","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"city":{"type":"string","description":"Free-text city or region (e.g. 'Limoges', 'Bay Area'). Resolved via the same /geo/search the followups_map uses. Ambiguous matches surface as `status: ambiguous_locations` with location_ambiguities[]; pick a location id and re-call with city_id. NEVER a country name — and unlike the Monitor tools the fix is NOT to omit this argument: a tour with no city returns arbitrary leads from the whole workspace, which is not an itinerary. Ask which city or region the user is visiting and pass that."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"city_id":{"type":"string","description":"Pre-resolved admin_area id (numeric string). Bypasses the resolver. Pass `city` alongside it with the NAME of the area you picked: the id scopes the Monitor half, and the name is the only thing that can scope the Discover half, which has no server-side geo filter. With `city_id` alone, `discover_leads` comes back empty."},"radius_km":{"type":"number","description":"How far outside the named town a Discover stop may be, in km. Default 20, which is roughly a metro area — it is what puts West Sacramento on a tour of Sacramento and Courbevoie on a tour of Paris. Pass the user's own number when they name one (\"dans un rayon de 10km\", \"zone 20km\", \"within 15 miles\" → 24). Pass 0 to keep the tour inside the town's own name. Leads in the town itself always come first; the radius only adds to them. It applies to Discover leads only — the Monitor half is scoped server-side and is untouched."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"discover_count":{"type":"number","description":"Top-N Discover leads (active lens wishlist) to return after client-side city filter. Default 6."},"followups_count":{"type":"number","description":"Top-N follow-up (Monitor) leads to return. Default 6."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["monitor_leads","discover_leads","map_locations"],"properties":{"city":{"type":["string","null"]},"_meta":{"type":"object","properties":{"region":{"type":"string"},"latency_ms":{"type":["number","null"]}}},"status":{"type":"string","description":"'ambiguous_locations' when the passed `city` matched multiple admin areas — pick an id from location_ambiguities and re-call with city_id. 'country_level_location' when `city` was a country name — do NOT drop the argument (a city-less tour is arbitrary nationwide leads); ask which city or region to use. The itinerary arrays are empty and nothing was fetched."},"city_id":{"type":["string","null"]},"next_steps":{"type":["object","null"],"properties":{"options":{"type":"array","items":{"type":"object","required":["label","description","kind"],"properties":{"kind":{"type":"string"},"label":{"type":"string"},"description":{"type":"string"}}}},"question":{"type":"string"}},"description":"Deterministic follow-on offers, the route planner FIRST. Map `options[]` into the host's next-step widget VERBATIM and in order — do not reword, reorder or drop them. null when the tour found no leads. Each option: {label (≤5 words), description (the full sentence), kind}."},"map_summary":{"type":"object","properties":{"total_leads":{"type":"number"},"leads_with_coords":{"type":"number"},"leads_without_coords":{"type":"number"}},"description":"Deterministic coverage counts so the agent can footnote '+ N leads without coordinates' without re-counting."},"map_locations":{"type":"array","items":{"type":"object"},"description":"Pre-shaped entries for `places_map_display_v0` — pass each one verbatim ({name, address, latitude, longitude, notes}); the mode badge (★ Customer / ★ Qualified / ✦ New) is already in `notes`. Do NOT reshape or re-derive from `location.pos`. One entry per lead with valid coordinates; coordinate-less leads are omitted and counted in `map_summary`."},"monitor_leads":{"type":"array","items":{"type":"object"},"description":"Follow-up (Monitor) leads in the requested city, sorted by AI / firmographic score. Each lead carries the same shape as pull_followups."},"discover_leads":{"type":"array","items":{"type":"object"},"description":"Fresh Discover leads from the active lens, filtered client-side to match the city, then extended with the ones whose own coordinates put them within `radius_km` of it. The town's own leads come first. Pulls a larger candidate set internally to compensate for the missing server-side geo filter."},"country_locations":{"type":"array","items":{"type":"object"},"description":"Per offending value: {value, param, kind, country}. Only present when `status === 'country_level_location'`. Unlike the Monitor tools, the recovery here is NOT to drop `city`: a tour with no city returns arbitrary leads from the whole workspace, which is not an itinerary. Ask which city or region the user is visiting and re-call with that — see `hint`."},"discover_filter_note":{"type":"string","description":"Human-readable summary of the client-side geo filter applied to Discover leads. Says how many stops are in the named town, how many are within `radius_km` of it, and which towns those are in — repeat that split to the user rather than presenting every stop as being in the city they named."},"location_ambiguities":{"type":"array","items":{"type":"object"}}}}},{"name":"leadbay_list_campaigns","annotations":{"title":"List your campaigns (with roll-up stats)","readOnlyHint":true,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$77","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"archived":{"type":"boolean","description":"Include archived campaigns only. Default false (active only)."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["campaigns"],"properties":{"_meta":{"type":"object","properties":{"region":{"type":"string"},"latency_ms":{"type":["number","null"]}}},"campaigns":{"type":"array","items":{"type":"object"},"description":"Each entry is {campaign, lead_count, contact_count, contacted, meeting_booked, declined}. `contacted` = leads with at least one logged outreach; `meeting_booked` = leads with a recorded meeting outcome; `declined` = leads with a recorded decline outcome."}}}},{"name":"leadbay_campaign_progression","annotations":{"title":"Read per-lead progression inside a campaign","readOnlyHint":true,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$78","inputSchema":{"type":"object","required":["campaign_id","_triggered_by"],"properties":{"page":{"type":"number","description":"0-indexed page (default 0)."},"count":{"type":"number","description":"Leads per page (default 50, server-capped)."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"campaign_id":{"type":"string","description":"Campaign UUID (from leadbay_create_campaign or leadbay_list_campaigns)."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["items","pagination","summary"],"properties":{"_meta":{"type":"object","properties":{"region":{"type":"string"},"latency_ms":{"type":["number","null"]}}},"items":{"type":"array","items":{"type":"object"},"description":"Per-lead progression rows: {lead, progress: {total_contacts, in_progress, declined, headline}, affiliation: {own_campaigns, other_users_campaign_count}}. `headline` is the most recent interaction type (e.g. CONTACTED, MEETING_BOOKED, DECLINED)."},"summary":{"type":"object","properties":{"declined":{"type":"number"},"contacted":{"type":"number"},"page_size":{"type":"number"},"in_progress":{"type":"number"}},"description":"Roll-up across the current page: how many leads have any outreach, how many converted to meetings, how many were declined."},"pagination":{"type":"object","properties":{"page":{"type":"number"},"pages":{"type":"number"},"total":{"type":"number"}}}}}},{"name":"leadbay_campaign_call_sheet","annotations":{"title":"Cold-calling cheat sheet for a campaign","readOnlyHint":true,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$79","inputSchema":{"type":"object","required":["campaign_id","_triggered_by"],"properties":{"page":{"type":"number","description":"0-indexed page (default 0)."},"count":{"type":"number","description":"Leads per page for the underlying /campaigns/{id}/leads fetch (default 50). The contacts endpoint is not paginated; the composite always fetches all contacts and aligns them with the leads page."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"campaign_id":{"type":"string","description":"Campaign UUID (from leadbay_create_campaign or leadbay_list_campaigns)."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["campaign_id","leads","summary"],"properties":{"_meta":{"type":"object","properties":{"region":{"type":"string"},"latency_ms":{"type":["number","null"]}}},"leads":{"type":"array","items":{"type":"object"},"description":"Per-lead call-ready blocks sorted by score desc. Each block: {lead_id, lead_name, score, ai_agent_lead_score, location, website, next_step, approach_angle, worth_pursuing, progress, affiliation, contacts: [{first_name, last_name, full_name, phone_number, phone_tel_url, email, mailto_url, linkedin_url, linkedin_url_source, job_title, recommended, pinned, recent_notes}]}."},"summary":{"type":"object","properties":{"total_leads":{"type":"number"},"total_contacts":{"type":"number"},"leads_with_email":{"type":"number"},"leads_with_phone":{"type":"number"},"leads_with_coords":{"type":"number"},"leads_without_contacts":{"type":"number"},"leads_already_contacted":{"type":"number"}},"description":"Roll-up: total_leads, total_contacts, leads_with_phone, leads_with_email, leads_with_coords, leads_without_contacts, leads_already_contacted (headline/in_progress/declined outreach signal present)."},"readiness":{"type":"object","properties":{"travel_friendly":{"type":"boolean"},"needs_enrichment":{"type":"boolean"},"ready_for_calling":{"type":"boolean"},"ready_for_emailing":{"type":"boolean"}},"description":"Pre-computed booleans the orchestrator prompt uses to decide which session modes to OFFER. ready_for_calling (phone coverage ≥60%), ready_for_emailing (email coverage ≥60%), needs_enrichment (≥30% no-contact leads OR both phone+email coverage <40%), travel_friendly (≥5 geocoded leads AND coord coverage ≥60%)."},"next_steps":{"type":["object","null"],"properties":{"options":{"type":"array","items":{"type":"object","required":["label","description","kind"],"properties":{"kind":{"type":"string"},"label":{"type":"string"},"description":{"type":"string"}}}},"question":{"type":"string"}},"description":"Deterministic follow-on offers, artifact option FIRST. Map `options[]` into the host's next-step widget VERBATIM and in order — do not reword, reorder or drop them. null when the sheet is empty. Each option: {label (≤5 words), description (the full sentence), kind}."},"pagination":{"type":"object","properties":{"page":{"type":"number"},"pages":{"type":"number"},"total":{"type":"number"}}},"campaign_id":{"type":"string"},"map_locations":{"type":"array","items":{"type":"object"},"description":"Ready-to-pass entries for `places_map_display_v0` when the user wants a geographic call route: {name, address, latitude, longitude, notes}. One entry per lead that has a valid pos[]. Skip leads without coordinates."}}}},{"name":"leadbay_research_lead_by_id","annotations":{"title":"Research a Leadbay lead in depth (by UUID)","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$7a","inputSchema":{"type":"object","required":["leadId","_triggered_by"],"properties":{"leadId":{"type":"string","description":"Full 36-character lead UUID exactly as returned in leadbay_pull_leads items[].id (never shortened)"},"lensId":{"type":"number","description":"Lens id (escape hatch — normally omit; auto-resolves to the active lens)"},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"concise":{"type":"boolean","description":"If true, trim signals to hot=true items only (smaller payload). Default false."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"response_format":{"enum":["json","markdown"],"type":"string","description":"How the agent wants the result rendered. 'json' (default): the structured payload as text. 'markdown': a compact human-readable rendering (sections + bullets) — useful for chat-rendering clients (Cursor, Claude Desktop) where the user sees the response directly. structuredContent is emitted in both modes so capable clients still get typed access."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["qualification","signals","firmographics","contacts","engagement","recent_activities"],"properties":{"_meta":{"type":"object","properties":{"region":{"type":"string"},"lens_id":{"type":"number"},"resolved_from":{"type":["string","null"]},"resolved_query":{"type":["string","null"]},"match_candidates":{"type":["array","null"],"items":{"type":"object"}},"resolved_matched_on":{"type":["array","null"],"items":{"type":"string"}},"has_reachable_contact":{"type":"boolean"},"web_fetch_in_progress":{"type":"boolean"}},"description":"Operator context: region (us/fr/custom), lens_id (the lens used for the lead-by-id fetch), web_fetch_in_progress (true if the backend is still hydrating signals), has_reachable_contact (true if at least one contact or recommended_contact has email or phone — drives NEXT STEPS routing between enrichment vs outreach). When the call was routed via leadbay_research_lead_by_name_fuzzy, also: resolved_from='companyName' (matched in the user's own leads) or 'resolver' (matched in the Leadbay company registry), resolved_query='', resolved_matched_on=['website_exact',…] on the resolver path, match_candidates=[{leadId,name,score}].","additionalProperties":true},"signals":{"type":"array","items":{"type":"object"},"description":"Web-research signals reshaped into priority-ordered sections (profile → signals → clues → other). Each entry: section_label, section_emoji, entries[]. With concise:true, only hot=true entries kept. May be auto-trimmed when truncated:true (see below)."},"contacts":{"type":"object","properties":{"reachable":{"type":"array","items":{"type":"object"}},"candidates":{"type":"array","items":{"type":"object"}}},"description":"$7b","additionalProperties":false},"truncated":{"type":"boolean","description":"True when the response was auto-trimmed to stay under the ~25k-char budget. Always false when concise:true is passed."},"engagement":{"type":"object","properties":{"new":{"type":"boolean"},"liked":{"type":"boolean"},"disliked":{"type":"boolean"},"notes_count":{"type":"number"},"recommended_contact":{"type":["object","null"]},"epilogue_actions_count":{"type":"number"},"prospecting_actions_count":{"type":"number"}},"description":"What humans/prior agent runs already did: liked/disliked flags, recommended_contact, and counts (notes/epilogue/prospecting). Recent items live in `recent_activities` at the top level.","additionalProperties":false},"firmographics":{"type":"object","properties":{"id":{"type":"string"},"name":{"type":"string"},"size":{"type":["object","null"],"properties":{"low":{"type":["number","null"]},"max":{"type":["number","null"]},"min":{"type":["number","null"]},"high":{"type":["number","null"]},"label":{"type":["string","null"]}},"description":"LeadSimplified.size — employee-count band."},"tags":{"type":"array","items":{"type":"object","properties":{"id":{"type":["number","string","null"]},"tag":{"type":"string"},"score":{"type":["number","null"]},"display_name":{"type":["string","null"]}}},"description":"LeadTag[] — {id, display_name, tag, score}."},"score":{"type":["number","null"]},"website":{"type":["string","null"]},"keywords":{"type":"array","items":{},"description":"Either bare strings (legacy) or {keyword,score} objects depending on the backend payload version."},"location":{"type":["object","null"],"properties":{"pos":{"type":["array","null"],"items":{"type":"number"}},"city":{"type":["string","null"]},"full":{"type":["string","null"]},"state":{"type":["string","null"]},"country":{"type":["string","null"]}},"description":"LeadFullLocation — city/state/country/full/pos."},"sector_id":{"type":["number","string","null"]},"ai_summary":{"type":["string","null"]},"description":{"type":["string","null"]},"social_urls":{"type":["object","null"],"properties":{"tiktok":{"type":["string","null"]},"twitter":{"type":["string","null"]},"facebook":{"type":["string","null"]},"linkedin":{"type":["string","null"]},"instagram":{"type":["string","null"]},"crunchbase":{"type":["string","null"]}},"description":"LeadSocialUrls — URL strings per platform; null when the company has no profile."},"registry_ids":{"type":["object","array","null"]},"phone_numbers":{"type":["array","null"],"items":{"type":"string"}},"social_presence":{"type":["object","null"],"properties":{"tiktok":{"type":"boolean"},"twitter":{"type":"boolean"},"facebook":{"type":"boolean"},"linkedin":{"type":"boolean"},"instagram":{"type":"boolean"},"crunchbase":{"type":"boolean"}},"description":"LeadSocialPresence — 6 booleans per platform. Use `social_urls` for URLs."},"split_ai_summary":{"type":["object","null"],"properties":{"next_step":{"type":["string","null"]},"approach_angle":{"type":["string","null"]},"worth_pursuing":{"type":["string","null"]}}},"short_description":{"type":["string","null"]},"ai_agent_lead_score":{"type":["number","null"]}},"description":"Lead profile basics. The shapes here match the backend `LeadSimplified` schema verbatim — `size` is `{min,max,...}`, `location` is `{city,state,country,full,pos}`, `tags` are typed objects.","additionalProperties":false},"qualification":{"type":"array","items":{"type":"object","properties":{"question":{"type":"string"},"response":{"type":["string","null"]},"boost_score":{"type":["number","null"]},"computed_at":{"type":["string","null"]},"score_scale":{"type":"string"},"score_0_to_10":{"type":["number","null"]}}},"description":"Per-question AI qualification answers, ordered by mission importance. Each entry: question, boost_score (canonical -10|0|10|20), score_scale, response, computed_at. score_0_to_10 is a deprecated alias of boost_score (removed in 0.7.0)."},"truncation_hint":{"type":["string","null"],"description":"When truncated:true, names the specific argument that would reduce the payload (typically 'concise:true')."},"recent_activities":{"type":"array","items":{"type":"object","properties":{"date":{"type":"string"},"type":{"type":"string"}}},"description":"Unified activity timeline (top 20 most recent) from /leads/{id}/activities. Each entry: {type, date}. Replaces the per-category recent_notes/recent_epilogue/recent_prospecting arrays from the prior schema — counts stay on `engagement` as the cheap signal."},"web_insights_fetched_at":{"type":["string","null"],"description":"ISO timestamp of the latest /web_fetch run. Use as a staleness signal — if older than 30 days, offer to refresh."}}}},{"name":"leadbay_research_lead_by_name_fuzzy","annotations":{"title":"Look up a lead by company name (fuzzy)","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$7c","inputSchema":{"type":"object","required":["companyName","_triggered_by"],"properties":{"email":{"type":"string","description":"A contact email at the company. Used to derive the company domain when `website` is absent; consumer mailboxes (gmail, orange.fr, …) are ignored."},"lensId":{"type":"number","description":"Optional strict scope. When supplied, search only this lens's wishlist and do NOT fall through to the registry; normally omit."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"concise":{"type":"boolean","description":"Forwarded to leadbay_research_lead_by_id. If true, trims signals to hot=true items only."},"website":{"type":"string","description":"Company domain or website when you have one (`acme.com`, `https://www.acme.com/` — both fine). This is the single strongest match key; pass it whenever the user mentioned a domain, and a company outside their leads becomes findable."},"companyName":{"type":"string","description":"Company name, domain, or contact name. Resolved against the user's own Discover/Monitor/Activate leads and the Leadbay company registry."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"registry_number":{"type":"string","description":"Company registry number (SIREN/SIRET in France, company number elsewhere). The other exact match key — pass it when the user supplies one, or when a previous LEAD_NOT_FOUND hint asked for it."},"response_format":{"enum":["json","markdown"],"type":"string","description":"Forwarded to leadbay_research_lead_by_id. Default 'json'."}},"additionalProperties":false},"outputSchema":{"type":"object","description":"Same shape as leadbay_research_lead_by_id, with _meta.resolved_from='companyName'|'resolver', _meta.resolved_query='', _meta.resolved_matched_on=[...], and _meta.match_candidates=[{leadId,name,score}] populated. When the registry resolver cannot pick one company, returns {resolution:'ambiguous', query, candidates:[{leadId,name,website,location,registry_ids,score,matched_on}]} instead — ask the user which one, then call leadbay_research_lead_by_id. When neither the user's leads nor the registry hold the company, returns {resolution:'not_found', query, summary, would_help, next_step} — a successful answer, not an error: do what next_step says, which is normally to ask the user for the field would_help names ('website' or 'registry_number') and call again. would_help is empty when the search was scoped to one lens, where the remedy is to drop the scope, not to ask the user for a field.","additionalProperties":true}},{"name":"leadbay_get_qualification_questions","annotations":{"title":"Read the org's qualification questions","readOnlyHint":true,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$7d","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["qualification_questions"],"properties":{"hint":{"type":"string","description":"Operator note — admin edit pointer, or the empty-state message when no questions are configured."},"_meta":{"type":"object"},"count":{"type":"number","description":"Number of qualification questions configured."},"region":{"type":"string"},"is_admin":{"type":"boolean","description":"Whether the current bearer-token holder is an org admin. Admins can modify the questions via leadbay_set_qualification_questions."},"targeting_prompt":{"type":["string","null"],"description":"The org's free-text targeting prompt (user_prompt) the AI agent follows, or null when unset. Qualitative rules live here rather than in a question; change it with leadbay_refine_lead_targeting."},"ideal_buyer_profile":{"description":"The org's Ideal Buyer Profile {summary, key_characteristics, anti_patterns}, or null when none is configured. The questions score against this profile — read it before judging whether a user's stated rule is already covered."},"qualification_questions":{"type":"array","items":{"type":"object"},"description":"Org-level questions Leadbay scores every lead against. Each: {question, created_at, lang}."}}}},{"name":"leadbay_getting_started","annotations":{"title":"Guided Leadbay walkthrough","readOnlyHint":true,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$7e","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["version","intro","steps","keep_going","stop"],"properties":{"stop":{"type":"string","description":"How to close the tour."},"_meta":{"type":"object","description":"Server-injected notices, when present."},"intro":{"type":"string","description":"What to say before gate 1."},"steps":{"type":"array","description":"The gates, in order. Each carries n, gate_label, gate_description, explain, next_steps (the widget payload), calls, args, and optional forbidden_args / spend / quota_note / pin / branches. Render next_steps VERBATIM."},"version":{"type":"number","description":"Manifest version; bump means the script changed."},"docs_url":{"type":"string","description":"The canonical setup guide."},"docs_note":{"type":"string","description":"The only two moments that link may appear."},"exit_offer":{"type":"string","description":"How and when to make that offer."},"keep_going":{"type":"array","description":"The hand-off cards: what the user should TYPE to get each thing back once the buttons are gone. Every phrase is lifted from a tool's own routing triggers."},"calendly_url":{"type":"string","description":"1:1 setup session, offered on exit only."},"one_option_rule":{"type":"string","description":"The invariant every gate obeys: one way forward plus one way out, never three."}}}},{"name":"leadbay_get_lead_custom_fields","annotations":{"title":"Read a lead's custom-field values","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$7f","inputSchema":{"type":"object","required":["leadId","_triggered_by"],"properties":{"leadId":{"type":"string","description":"Lead UUID (required)"},"lensId":{"type":"number","description":"Lens id (escape hatch — normally omit; auto-resolves to the active lens)"},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["custom_fields","lead_id"],"properties":{"hint":{"type":"string","description":"Operator note — empty-state guidance, or a degradation note when the lead fetch returned entries without embedded definitions."},"_meta":{"type":"object"},"count":{"type":"number"},"region":{"type":"string"},"lead_id":{"type":"string"},"custom_fields":{"type":"array","items":{"type":"object"},"description":"Custom-field VALUES on this lead. Each: {id, name, type, value}. Empty when the org has no custom fields or none are set on this lead."}}}},{"name":"leadbay_account_history","annotations":{"title":"One account's full back-story","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$80","inputSchema":{"type":"object","required":["leadId","_triggered_by"],"properties":{"leadId":{"type":"string","description":"Lead UUID (required)"},"lensId":{"type":"number","description":"Lens id the lead came from (escape hatch — normally omit; defaults to the active lens). Pass it when researching a lead from a lens other than the current default, so the underlying /lenses/{lensId}/leads/{leadId} fetch doesn't 404 after the active lens changed."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"activityCount":{"type":"number","description":"Number of activity-timeline entries to return, max 100 (default: 50)."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["lead","notes","activities"],"properties":{"lead":{"type":"object","required":["id"],"properties":{"id":{"type":"string"},"name":{"type":["string","null"]}},"description":"Which account this history is about."},"_meta":{"type":"object","description":"Research's pass-through metadata (lens_id, web_fetch_in_progress, has_reachable_contact, …) plus region and this call's counts."},"notes":{"type":"array","description":"FULL note bodies — the part research only counts. Oldest first."},"signals":{"type":["array","null"],"description":"Live signals, verbatim from leadbay_research_lead_by_id — why this account is hot NOW."},"contacts":{"type":["object","null"],"description":"Reachable contacts + candidates, verbatim from research."},"activities":{"type":"object","required":["activities","total"],"properties":{"total":{"type":"number","description":"Total on the server, which can exceed the number returned."},"activities":{"type":"array","items":{"type":"object","properties":{"date":{"type":"string"},"type":{"type":"string"}}},"description":"One entry per logged interaction."}},"description":"The interaction timeline research only summarizes."},"engagement":{"type":["object","null"],"description":"Engagement counters, verbatim from research."},"firmographics":{"type":["object","null"],"description":"Company facts, verbatim from research."},"qualification":{"type":"array","description":"Per-question qualification answers, verbatim from research."}}}},{"name":"leadbay_scan_portfolio_signals","annotations":{"title":"Scan a portfolio for a web-research signal in bulk","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$81","inputSchema":{"type":"object","required":["query","_triggered_by"],"properties":{"city":{"type":"string","description":"Free-text city / region to scope the Monitor portfolio before scanning (resolved via /geo/search, same as leadbay_pull_followups). Ignored when `leadIds` is given. NEVER a country name: this workspace serves exactly ONE country, so a whole-country ask means omitting `city` entirely."},"query":{"type":"string","description":"Signal terms to match (case- and accent-insensitive). Comma- or space-separated terms are OR'd, e.g. 'M&A, acquisition, racheté'. Matched against each signal entry's description, source, and section label."},"since":{"type":"string","description":"ISO date (e.g. '2025-01-01'). When set, only signal entries dated on/after it are returned. Entries with no date are kept (absence of a date is not evidence the event is old)."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"city_id":{"type":"string","description":"Pre-resolved admin_area id (numeric string). Bypasses the resolver. Ignored when `leadIds` is given."},"leadIds":{"type":"array","items":{"type":"string"},"description":"Explicit lead UUIDs to scan (skips Monitor pagination). Use when you already hold a cohort of ids."},"max_leads":{"type":"number","description":"Cap on leads scanned (default 200, hard max 300). When the portfolio exceeds this, the scan is truncated and truncated_at is set."},"set_filter":{"type":"object","properties":{"criteria":{"type":"array","items":{"type":"object"}}},"description":"Optional Monitor FilterItem ({criteria: FilterCriterion[]}) to scope the portfolio before scanning. Persisted server-side then applied, mirroring leadbay_pull_followups. Ignored when `leadIds` is given. A `location_ids` criterion must carry sub-country admin areas only — a country name here is rejected with COUNTRY_LEVEL_LOCATION before anything is persisted."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["matched","not_researched","scanned_count","matched_count","quota_exceeded"],"properties":{"_meta":{"type":"object","properties":{"region":{"type":"string"}}},"status":{"type":"string","description":"`ambiguous_locations` when a passed `city` matched multiple admin_areas; pick an id from `location_ambiguities` and re-call with `city_id`. `country_level_location` when `city`, `city_id` or a `set_filter` `location_ids` criterion carried a country name — nothing was scanned and no filter was persisted. Absent on the happy path."},"matched":{"type":"array","items":{"type":"object"},"description":"Leads with ≥1 signal entry matching the query. Each: {lead_id, name, location, matched_signals:[{section_label, section_emoji, hot, source, date, description}]}. Campaign-ready — feed lead_ids straight into leadbay_add_leads_to_campaign."},"truncated_at":{"type":"number","description":"Present only when the portfolio exceeded `max_leads`; equals the cap applied. Coverage is partial — narrow the scope (city / set_filter) or raise max_leads."},"matched_count":{"type":"number","description":"Length of `matched`."},"scanned_count":{"type":"number","description":"Total leads read in this scan (matched + non-matching + not_researched)."},"not_researched":{"type":"array","items":{"type":"object"},"description":"Leads scanned that had NO cached signal content (web_fetch.content null or still in progress). These are NOT 'no match' — they were never researched. Qualify them (leadbay_bulk_qualify_leads) then re-scan. Each: {lead_id, name}."},"quota_exceeded":{"type":"boolean","description":"True if a 429 was hit mid-scan. Partial `matched` is still returned. Offer wait-for-reset OR top-up."},"country_locations":{"type":"array","items":{"type":"object"},"description":"Per offending value: {value, param, kind, country, axis, kept}. Only present when `status === 'country_level_location'`. The recovery BRANCHES on `country_locations[].axis` and `[].kind`; `hint` states the one for THIS call — follow it verbatim. Dropping the argument is NOT the general answer: on an `exclude` axis it returns the very companies the user asked to remove, and for a `foreign_country` an unfiltered result is this workspace's own leads, which answer a different question. Never retry with another spelling or a nearby city."},"location_ambiguities":{"type":"array","items":{"type":"object"},"description":"Only present when status === 'ambiguous_locations'."}}}},{"name":"leadbay_team_activity","annotations":{"title":"Team activity + per-rep KPIs","readOnlyHint":true,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$82","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"to":{"type":"string","description":"ISO date YYYY-MM-DD (end). Overrides weeks."},"from":{"type":"string","description":"ISO date YYYY-MM-DD (start). Overrides weeks."},"weeks":{"type":"number","description":"Look-back window in weeks (default 4). Ignored if from/to given."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"user_ids":{"type":"array","items":{"type":"string"},"description":"Specific rep user-ids; omit for the whole team."},"periodicity":{"enum":["DAILY","WEEKLY"],"type":"string","description":"Trend bucketing (default WEEKLY)."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["range","reps","trend"],"properties":{"reps":{"type":"array","items":{"type":"object","properties":{"lost":{"type":"number"},"name":{"type":["string","null"]},"email":{"type":["string","null"]},"likes":{"type":"number"},"notes":{"type":"number"},"saves":{"type":"number"},"user_id":{"type":"string"},"exported":{"type":"number"},"profile_views":{"type":"number"},"still_chasing":{"type":"number"},"contacts_added":{"type":"number"},"website_clicks":{"type":"number"},"could_not_reach":{"type":"number"},"total_activities":{"type":"number"},"contacts_purchased":{"type":"number"},"meetings_or_interest":{"type":"number","description":"Epilogue outcomes logged in the window — the 'deals' signal."}}},"description":"One row per rep, already renamed from the backend's KPI vocabulary. Non-admins get only themselves; the backend does that scoping."},"_meta":{"type":"object","description":"Region + any server-injected notices."},"range":{"type":"object","required":["from","to","periodicity"],"properties":{"to":{"type":"string","description":"ISO date (YYYY-MM-DD), inclusive."},"from":{"type":"string","description":"ISO date (YYYY-MM-DD), inclusive."},"periodicity":{"type":"string","description":"DAILY or WEEKLY — the trend bucket size."}},"description":"The window actually queried, after `weeks` was resolved to dates."},"trend":{"type":"array","items":{"type":"object","properties":{"date":{"type":"string"},"count":{"type":"number"}}},"description":"Activity over time, one point per period."}}}},{"name":"leadbay_list_previously_enriched_titles","annotations":{"title":"Recall titles previously enriched","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"Show job titles the org has previously enriched, so the agent can repeat the same titles for new leads (or skip already-saturated ones). Two implementation paths: (1) PRIMARY — a selection-scoped preview call that reads `previously_enriched_titles` from the backend (newer prod field). (2) FALLBACK — live aggregation across each lead's enriched contacts. The composite picks transparently.\n\nWHEN TO USE: before leadbay_enrich_titles, to plan which titles to order.\n\nWHEN NOT TO USE: when you already know the exact titles you want to enrich.\n","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"lensId":{"type":"number","description":"Override the auto-resolved last-active lens when omitting leadIds (escape hatch)"},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"leadIds":{"type":"array","items":{"type":"string"},"description":"Lead UUIDs to query. Omit to use the current wishlist (top page)."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["source","titles"],"properties":{"note":{"type":"string","description":"Operator note explaining the chosen path (e.g., empty input set, fallback in use)."},"source":{"type":"string","description":"'preview_field' (preferred backend path) or 'live_aggregate' (fallback aggregation across each lead's contacts)."},"titles":{"type":"array","items":{"type":"object"},"description":"Titles previously enriched. preview_field path: [{title}]. live_aggregate path: [{title, leads_with_enriched, total_enriched_contacts, leads_still_having_unenriched}]."},"available_in_selection":{"type":"array","items":{"type":"object"},"description":"Backend-suggested title candidates available for ordering (preview_field path only)."}}}},{"name":"leadbay_account_status","annotations":{"title":"Show Leadbay account + quota state","readOnlyHint":true,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$83","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["user","organization"],"properties":{"user":{"type":"object","properties":{"name":{"type":["string","null"]},"admin":{"type":"boolean"},"email":{"type":["string","null"]},"manager":{"type":"boolean"},"language":{"type":"string"}},"description":"Identity & roles for the current bearer-token holder."},"_meta":{"type":"object","properties":{"region":{"type":"string"}}},"quota":{"type":["object","null"],"description":"$84"},"mcp_version":{"type":"string","description":"Version of the Leadbay MCP server answering this call. Answer 'what version of Leadbay are you running' with this value."},"quota_error":{"type":["object","null"],"properties":{"code":{"type":"string"},"message":{"type":"string"},"http_status":{"type":["number","null"]}},"description":"Non-null ONLY when the quota_status call FAILED — {code, http_status, message}. This is an INTERNAL diagnostic — do NOT surface it to the user AT ALL. A 401/403 here does NOT mean the login is broken: the `user`/`organization` fields in THIS SAME response were fetched with the same token and succeeded. It's a backend-side quirk (commonly an org with no billing plan — `plan: null`), irrelevant to the user. When quota is unreadable: stay SILENT — do NOT mention quota, do NOT say 'quota unreadable', do NOT mention any error or 401, and NEVER tell the user to reconnect or re-authenticate. Just answer the rest (user / org / lens) and omit quota entirely. NEVER report it as zero usage or 'no limits' either."},"organization":{"type":"object","properties":{"id":{"type":"string"},"name":{"type":"string"},"plan":{"type":["string","null"]},"ai_agent_enabled":{"type":"boolean"},"unlimited_credits":{"type":"boolean","description":"True for internal/unlimited accounts (billing disabled server-side). When true, DO NOT read the null `plan` or absent quota as 'no credits' — the account has unlimited credits; enrichment/qualify proceed freely. Stay SILENT on credits: don't announce 'unlimited', don't show a balance, and never tell the user they're out of credits. False for normal accounts, where the usual quota/credit story applies."},"computing_intelligence":{"type":"boolean","description":"True if Leadbay is mid-regenerating intelligence after a targeting refinement; new leads will reflect it shortly."}},"description":"Org-level state and feature flags."},"notifications":{"type":"array","items":{"type":"object"},"description":"Terminal bulk-progress notifications the MCP knows about (background work the user or agent started that has since completed). Each entry carries notification_id, kind (bulk_enrich | bulk_qualify | import | other), bulk_progress counters, and a revise_hint pointing at prior agent outputs the just-finished work might have made stale. After revising affected outputs, call leadbay_acknowledge_notification(notification_id) to clear the entry. Empty array when nothing has completed."},"update_available":{"type":["object","null"],"required":["current_version","latest_version","install_url","release_url"],"properties":{"install_url":{"type":"string","description":"Direct download URL for the installer asset (.dxt, falling back to .mcpb)."},"release_url":{"type":"string","description":"GitHub release page (changelog)."},"latest_version":{"type":"string"},"current_version":{"type":"string"}}},"last_requested_lens":{"type":["string","null"],"description":"The most recent lens id (a STRING, e.g. \"40005\"). WITHHELD (null) unless the user's message asked about the lens/audience — the composite only populates it when asked, so a plain 'what account?' answer has no lens to show. Even when present, this raw id is internal routing only — NEVER show the number to the user; use `last_requested_lens_name`."},"search_configuration":{"type":"object","properties":{"targeting_prompt":{"type":["string","null"],"description":"The org's free-text targeting prompt, in the user's own words. Changed with leadbay_refine_lead_targeting."},"ideal_buyer_profile":{"type":["object","null"],"properties":{"summary":{"type":"string"},"anti_patterns":{"type":"array","items":{"type":"string"}},"key_characteristics":{"type":"array","items":{"type":"string"}}},"description":"The org's Ideal Buyer Profile: who Leadbay looks for."},"purchase_intent_tags":{"type":["array","null"],"items":{"type":"string"},"description":"The buying signals Leadbay watches for, by name."},"qualification_questions":{"type":["array","null"],"items":{"type":"string"},"description":"The questions Leadbay scores every lead against. Changed with leadbay_set_qualification_questions."}},"description":"What this account is configured to search for — the org's targeting settings. A null field is unset OR could not be read: say nothing about it, and never tell the user it is missing. An empty list means none are configured."},"last_requested_lens_name":{"type":["string","null"],"description":"Human-readable name of the active lens (resolved from /lenses). WITHHELD (null) unless the user explicitly asked about the lens/audience — the composite only resolves it when asked, so on a plain account question there is nothing here to mention. When present (the user asked), answer with THIS name, never the numeric id."}}}},{"name":"leadbay_bulk_enrich_status","annotations":{"title":"Poll bulk-enrichment status","readOnlyHint":true,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$85","inputSchema":{"type":"object","anyOf":[{"required":["notification_id"]},{"required":["lead_ids"]}],"required":["_triggered_by"],"properties":{"email":{"type":"boolean","description":"The `email` flag the launch returned. Scopes counting to reservations that asked for email (`enrichment.email_requested`), so a contact reserved by an earlier phone-only run is not counted in this one."},"phone":{"type":"boolean","description":"The `phone` flag the launch returned. Same rule as `email`, on `enrichment.phone_requested`."},"titles":{"type":"array","items":{"type":"string"},"description":"The `titles` the launch returned. Scopes progress to the roles THIS run enriched, so a lead's pre-existing CFO email cannot inflate a CEO run."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"lead_ids":{"type":"array","items":{"type":"string"},"description":"The `lead_ids` the launch returned. Gives per-lead progress, and answers on its own if the notification is archived or has aged off page 1."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"notification_id":{"type":"string","description":"The `notification_id` returned by leadbay_enrich_titles. Answers on its own: the tool reads the job's lead set back from the backend, so an id kept from an earlier conversation still gives per-lead progress."},"include_contacts":{"type":"boolean","description":"If true, return the full contact list per lead (email, phone, enrichment.done). Default false — cheap status polls."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["status","leads","overall_progress","all_done"],"properties":{"status":{"type":"string","description":"'launched' on success. Errors return error envelopes (handled separately)."},"all_done":{"type":"boolean","description":"True when overall_progress.done === total AND no partial_failures."},"launched_at":{"type":"string","description":"ISO timestamp of /enrichment/launch ack."},"notification_id":{"type":"string","description":"The backend job id; pass it back to poll again."},"overall_progress":{"type":"object","properties":{"done":{"type":"number"},"total":{"type":"number"},"done_ratio":{"type":"number"}},"description":"Aggregate progress across all leads."},"partial_failures":{"type":"array","items":{"type":"object"},"description":"Per-lead errors observed during contacts fan-out (omitted when no failures)."},"credits_remaining":{"type":["number","string","null"],"description":"Advisory internal context only — do NOT display it. It's billing.ai_credits (a CONSUMED counter, not a remaining balance), so it can read 0 on a fresh/quota-backed account and printing _(N credits remaining)_ would falsely say 'out of credits'. Enrichment is gated by QUOTA, not this number — never a credits line here. (\"unlimited\" = internal/unlimited account; still nothing to display.) A per-run 'credits used' figure is intentionally NOT returned — getContacts can't scope cost to this bulk."}}}},{"name":"leadbay_qualify_status","annotations":{"title":"Poll import-and-qualify status","readOnlyHint":true,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$86","inputSchema":{"type":"object","anyOf":[{"required":["notification_id"]},{"required":["lead_ids"]}],"required":["_triggered_by"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"lens_id":{"type":"number","description":"The `lens_id` the launch returned. Used to flag leads no longer in the lens."},"lead_ids":{"type":"array","items":{"type":"string"},"description":"The `lead_ids` the launch returned. Supply them for per-lead detail (which settled, which are still running). Progress alone needs only notification_id."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"notification_id":{"type":"string","description":"The `notification_id` returned by leadbay_import_and_qualify / leadbay_bulk_qualify_leads. Answers progress in ONE call."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["notification_id","status","import_ids","lens_id","lead_ids","qualified","still_running","failed","not_in_lens","region","_meta"],"properties":{"_meta":{"type":"object"},"failed":{"type":"array","items":{"type":"object"},"description":"Per-lead errors observed at refresh (e.g., 404 on /web_fetch + /ai_agent_responses)."},"region":{"type":"string"},"status":{"type":"string","description":"'launched' on success (other states surface as error envelopes)."},"lens_id":{"type":"number","description":"Lens id the qualification ran against."},"lead_ids":{"type":"array","items":{"type":"string"},"description":"Lead UUIDs covered by this status (echoed from launch)."},"qualified":{"type":"array","items":{"type":"object"},"description":"Leads whose qualification has settled. Each entry: {lead_id, qualification_summary, signals_count, ...}."},"import_ids":{"type":"array","items":{"type":"string"},"description":"Underlying file-import handle ids (one per chunk)."},"launched_at":{"type":"string","description":"ISO timestamp of original launch."},"not_in_lens":{"type":"array","items":{"type":"string"},"description":"Lead ids that exist in the org but aren't members of the active lens — backend won't qualify them; agent should stop polling."},"still_running":{"type":"array","items":{"type":"object"},"description":"Leads still being qualified at refresh time."},"notification_id":{"type":["string","null"],"description":"The backend job id this status is for; pass it back to poll again."},"total_budget_ms":{"type":"number","description":"Caller-supplied total timeout (informational only at status time)."},"per_lead_budget_ms":{"type":"number","description":"Caller-supplied per-lead timeout (informational only at status time)."}}}},{"name":"leadbay_import_status","annotations":{"title":"Poll import status","readOnlyHint":true,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$87","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"dry_run":{"type":"boolean","description":"Pass true when the importIds came from a dry run. A dry run and an import still committing its mappings look identical on the wire, so without this the tool reports the dry run as still running rather than risk rendering it as a real import."},"importIds":{"type":"array","items":{"type":"string"},"description":"Backend file-import ids to inspect directly — from a completed import's `importIds`, or from a `{status:'running', timed_out:true}` result."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["status","importIds","progress","region","_meta"],"properties":{"_meta":{"type":"object"},"error":{"type":"string"},"region":{"type":"string"},"result":{"type":"object","description":"Final import result: {leads, not_imported, importIds, still_settling?}. Present when the importIds[] path finds every import complete and reconciles the wizard's records."},"status":{"type":"string","description":"running, complete, or failed."},"dry_run":{"type":"boolean","description":"True when these importIds came from a dry run. NOTHING was committed to the CRM — render it as a validation pass, never as a completed import."},"progress":{"type":"object"},"importIds":{"type":"array","items":{"type":"string"}}}}},{"name":"leadbay_resolve_import_rows","annotations":{"title":"Resolve import row identities","readOnlyHint":true,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$88","inputSchema":{"type":"object","required":["records","_triggered_by"],"properties":{"lensId":{"type":"number","description":"Lens ID used for candidate profile hydration. Defaults to the user's active lens."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"records":{"type":"array","items":{"type":"object"},"description":"CSV-shaped rows from the user file. Values may be strings, numbers, booleans, null, arrays, or objects; non-scalars are JSON-stringified for resolver/import preparation."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"identity_mappings":{"type":"object","properties":{"city":{"type":"string"},"name":{"type":"string"},"email":{"type":"string"},"phone":{"type":"string"},"crm_id":{"type":"string"},"tiktok":{"type":"string"},"address":{"type":"string"},"country":{"type":"string"},"twitter":{"type":"string"},"website":{"type":"string"},"facebook":{"type":"string"},"linkedin":{"type":"string"},"postcode":{"type":"string"},"instagram":{"type":"string"},"leadbay_id":{"type":"string"},"registry_type":{"type":"string"},"registry_number":{"type":"string"}},"description":"Resolver field -> source column map chosen by the agent after inspecting the file, e.g. {website:'company_domain', name:'Company', crm_id:'Salesforce ID', registry_number:'SIREN'}. May point to clean columns the agent derived before calling this tool. This tool does not infer mappings from header names.","additionalProperties":false},"candidate_profile_limit":{"type":"number","description":"Maximum candidates to hydrate per ambiguous row when include_candidate_profiles=true (default 5)."},"include_candidate_profiles":{"type":"boolean","description":"When true, hydrate ambiguous candidate IDs with lightweight lead facts from the active lens. Use on small batches or rerun on only ambiguous rows; large ambiguous files can return many candidates."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["rows","records_for_import","mappings_for_import","identity_mappings_used","mapping_guidance","disambiguation_policy","summary","next_action","region","_meta"],"properties":{"rows":{"type":"array","items":{"type":"object"},"description":"Per-input resolution result. Matched rows include lead_id; ambiguous rows include candidates; none/unidentifiable rows explain what extra signal would help."},"_meta":{"type":"object"},"region":{"type":"string"},"summary":{"type":"object"},"next_action":{"type":"string"},"mapping_guidance":{"type":"array","items":{"type":"string"},"description":"Instructions for building the final import mappings from the source columns."},"records_for_import":{"type":"array","items":{"type":"object"},"description":"Import-ready records. Matched rows include LEADBAY_ID. Ambiguous/unresolved rows preserve user data and rely on website/name/CRM fields for normal or late matching."},"mappings_for_import":{"type":"object","description":"Safe identity-only mapping starter. The agent should review/extend this using mapping_guidance and leadbay_list_mappable_fields before importing."},"disambiguation_policy":{"type":"array","items":{"type":"string"},"description":"Rules the agent should follow before writing LEADBAY_ID onto ambiguous rows."},"identity_mappings_used":{"type":"object"}}}},{"name":"leadbay_list_lens_seed_candidates","annotations":{"title":"List candidate seeds for a lens extra-refill","readOnlyHint":true,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$89","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"limit":{"type":"number","description":"Max candidates to return, 1–50 (backend default 20)."},"lensId":{"type":"number","description":"Lens to fetch candidates for. Defaults to the user's last-active lens."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["lens","candidates"],"properties":{"lens":{"type":"object","properties":{"id":{"type":"number"}}},"candidates":{"type":"array","items":{"type":"object"},"description":"Ranked candidate leads — each is a valid seed for leadbay_extend_lens. Pick 3–5 that represent the kind of leads the user wants more of."}}}},{"name":"leadbay_list_mappable_fields","annotations":{"title":"List CRM-import mappable fields","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":false},"description":"$8a","inputSchema":{"type":"object","properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"for_records":{"type":"array","items":{"type":"object"},"description":"Optional sample of CSV-shaped rows (objects). When provided, the tool runs the wizard's preprocess on the first 50 rows and attaches mapping_hints + custom_field_candidates to the response."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["standard_fields","custom_fields","region","_meta"],"properties":{"_meta":{"type":"object"},"notes":{"type":"array","items":{"type":"string"},"description":"Operator notes (e.g., preprocess timeout, sample-size truncation)."},"region":{"type":"string"},"sample_rows":{"type":"array","items":{"type":"object"},"description":"First few rows of the preprocessed sample (only when for_records was passed)."},"custom_fields":{"type":"array","items":{"type":"object"},"description":"Org-defined custom fields. Each: {id, name, type, description, mapping_value:'CUSTOM.'}."},"mapping_hints":{"type":"array","items":{"type":"object"},"description":"Per-column AI-confidence suggestions (only when for_records was passed). Each: {column, target, confidence, reason}."},"standard_fields":{"type":"array","items":{"type":"object"},"description":"Built-in StandardCrmFieldType entries (LEAD_NAME, LEAD_WEBSITE, contact + location + sector). Each: {name, description, mapping_value}."},"custom_field_candidates":{"type":"array","items":{"type":"object"},"description":"Custom fields matching unmapped columns by exact / case-insensitive / fuzzy name (only when for_records was passed)."}}}},{"name":"leadbay_list_sectors","annotations":{"title":"List sector taxonomy","readOnlyHint":true,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"List the sector taxonomy (id + display name in the requested language). Default: `lang` follows the caller's language; `includeInvisible=false` returns ~1,091 visible sectors.\n\nWHEN TO USE: to resolve a free-text sector name (e.g. \"Healthcare\") into the sector ids that leadbay_adjust_audience needs.\n\nWHEN NOT TO USE: when you already have sector ids — pass them directly.\n","inputSchema":{"type":"object","properties":{"lang":{"type":"string","description":"BCP-47 language tag (default: en)"},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."},"includeInvisible":{"type":"boolean","description":"Include sectors hidden from the UI (default false; ~91k items if true)"}},"additionalProperties":false}},{"name":"leadbay_list_locations","annotations":{"title":"Search the geo / admin-area taxonomy","readOnlyHint":true,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$8b","inputSchema":{"type":"object","required":["q"],"properties":{"q":{"type":"string","description":"Free-text city / region name (e.g. 'Berlin', 'NYC', 'São Paulo'). Returns top-10 prefix matches sorted by relevance, each with an admin_area id usable in FilterCriterion.location_ids. A COUNTRY name is refused — the index holds no country nodes, so the lookup could only return a same-named town."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["results","parents"],"properties":{"status":{"type":"string","description":"`country_level_location` when `q` was a country name — `results` is empty on purpose. This workspace serves exactly ONE country, so there is no country to look up and no id to pass on. Absent on the happy path."},"parents":{"type":"array","items":{"type":"object"},"description":"Parent admin areas referenced by `results[].parent_ids`, returned for breadcrumb / hover-disambiguation rendering."},"results":{"type":"array","items":{"type":"object"},"description":"Matches sorted by relevance. Each entry: {id, country, level, name, parent_ids}. `level` is admin depth (5=region, 6=county, 7=township-area, 8=city/town)."},"country_locations":{"type":"array","items":{"type":"object"},"description":"Per offending value: {value, param, kind, country, axis, kept}. Only present when `status === 'country_level_location'`. Unlike the lead-reading tools, the recovery here is NOT to drop `q` and re-call: `q` is required and an empty lookup returns no results, not workspace-wide coverage. There is simply no country id to hand out — see `hint`."}}}},{"name":"leadbay_create_topup_link","annotations":{"title":"Generate Stripe checkout URL for AI-credits top-up","readOnlyHint":false,"openWorldHint":true,"idempotentHint":false,"destructiveHint":false},"description":"$8c","inputSchema":{"type":"object","properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["url"],"properties":{"url":{"type":"string","description":"Stripe-hosted checkout URL. Surface as a clickable link; the user completes payment in their browser. Top-up clears the throttle immediately after Stripe webhook lands."}}}},{"name":"leadbay_open_billing_portal","annotations":{"title":"Generate Stripe customer-portal URL for subscription management","readOnlyHint":false,"openWorldHint":true,"idempotentHint":false,"destructiveHint":false},"description":"Generate a one-shot Stripe customer-portal URL. Wraps `GET /1.6/stripe/portal` → `{url}`. Returns a fresh Stripe-hosted URL the user can open to manage their existing Leadbay subscription: change plan tier, swap payment method, view invoices. The agent does NOT make subscription changes itself — it surfaces the URL and lets the user act.\n\nSibling of `leadbay_create_topup_link`. Use cases differ:\n\n- **Top-up** (`leadbay_create_topup_link`) = one-shot purchase of AI credits to clear an immediate quota throttle. Doesn't change the plan.\n- **Billing portal** (this tool) = ongoing subscription management. Plan upgrades persist; they raise the recurring window caps (daily/weekly/monthly) rather than buying a one-time bucket.\n\nWHEN TO USE: when the user asks to upgrade / downgrade / change plan; manage payment methods; view invoices; or \"open my billing\".\n\nWHEN NOT TO USE: for one-off credit top-ups (use `leadbay_create_topup_link`); for read-only quota state (use `leadbay_account_status` / `leadbay_get_quota`).\n","inputSchema":{"type":"object","properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["url"],"properties":{"url":{"type":"string","description":"Stripe customer-portal URL. Surface as a clickable link; the user manages subscription / payment methods / invoices in their browser."}}}},{"name":"leadbay_prepare_outreach","annotations":{"title":"Prepare outreach package for a lead","readOnlyHint":false,"openWorldHint":true,"idempotentHint":false,"destructiveHint":false},"description":"$8d","inputSchema":{"type":"object","required":["leadId","_triggered_by"],"properties":{"enrich":{"type":"boolean","description":"If true and credits available, trigger enrichment on the recommended contact (default: false). Enrichment is async; re-call this tool (no enrich) after ~60s and check enrichment.complete to see if email/phone landed (B13: self-polling)."},"leadId":{"type":"string","description":"Lead UUID (required)"},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["recommended_contact","additional_contacts_count","total_contacts_count","enrichment","history"],"properties":{"lead":{"type":["object","null"],"properties":{"id":{"type":"string"},"name":{"type":"string"},"size":{"type":["object","null"]},"score":{"type":["number","null"]},"website":{"type":["string","null"]},"location":{"type":["object","null"]},"ai_summary":{"type":["string","null"]},"description":{"type":["string","null"]},"social_urls":{"type":["object","null"]},"phone_numbers":{"type":["array","null"],"items":{"type":"string"}},"social_presence":{"type":["object","null"]},"split_ai_summary":{"type":["object","null"]},"short_description":{"type":["string","null"]},"ai_agent_lead_score":{"type":["number","null"]}},"description":"Lead context for the brief. Expanded per B15: score, ai_summary, split_ai_summary, location, size, phone_numbers, website, description."},"_meta":{"type":"object","description":"Operator context: agent memory summary when enabled."},"history":{"type":"object","required":["notes","activities","activities_total"],"properties":{"notes":{"type":"array","items":{"type":"object","properties":{"date":{"type":"string"},"note":{"type":"string"},"author":{"type":["string","null"]},"contact":{"type":["string","null"]}}},"description":"Every note on the lead and on each of its org contacts, newest first: {date, author, contact, note}. `contact` names the person a note is about, null for a note on the company."},"activities":{"type":"array","items":{"type":"object","properties":{"date":{"type":"string"},"type":{"type":"string"}}},"description":"The 20 most recent timeline entries, newest first: {type, date}."},"activities_total":{"type":"number","description":"Timeline entries on the server, which can exceed the number returned."}},"description":"What the team already did on this lead and its people."},"signals":{"type":["array","null"],"items":{"type":"object"},"description":"Web-research signals, the same priority-ordered sections leadbay_research_lead_by_id returns ({section_label, section_emoji, entries[]}). Empty when the lead was never researched or its research could not be read; null when the profile could not be read."},"enrichment":{"type":"object","properties":{"hint":{"type":["string","null"]},"error":{"type":["string","null"]},"complete":{"type":"boolean"},"triggered":{"type":"boolean"}},"description":"Self-polling status (B13): triggered = whether this call kicked off enrichment; complete = whether the recommended contact now has email OR phone."},"qualification":{"type":["array","null"],"items":{"type":"object"},"description":"The lead's AI qualification answers ({question, score, response, computed_at, outdated_at}). Empty when none were computed; null when they could not be read."},"recommended_contact":{"type":["object","null"],"properties":{"email":{"type":["string","null"]},"job_title":{"type":["string","null"]},"last_name":{"type":["string","null"]},"contact_id":{"type":["string","null"]},"first_name":{"type":["string","null"]},"phone_number":{"type":["string","null"]},"linkedin_page":{"type":["string","null"]},"is_org_contact":{"type":["boolean","null"]}},"description":"Best contact to outreach to. Always returned in the post-enrichment shape (B21) — first_name/last_name/contact_id/email/phone_number/linkedin_page/job_title — with nulls in fields that aren't yet enriched."},"other_contacts_count":{"type":"number","description":"DEPRECATED: alias for additional_contacts_count. Will be removed in 0.10.0."},"total_contacts_count":{"type":"number","description":"Total contacts on this lead (recommended + others)."},"additional_contacts_count":{"type":"number","description":"How many other contacts exist beyond the recommended one (renamed from other_contacts_count per B16; both shipped for one release)."}}}},{"name":"leadbay_report_friction","annotations":{"title":"Report a problem to the Leadbay team","readOnlyHint":false,"openWorldHint":true,"idempotentHint":false,"destructiveHint":false},"description":"$8e","inputSchema":{"type":"object","required":["category","message","_triggered_by"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"message":{"type":"string","description":"What the user wants to report, in their own words (cap 500 chars). Required. If the user already stated the problem when asking you to report it, those words ARE the message — send them in the same turn; do NOT ask them to re-confirm wording they just gave you. Only go back to them when you would otherwise have to invent the wording. Never call this tool unprompted."},"category":{"enum":["silent_failure","repeated_request","wrong_result","dissatisfaction","missing_capability","other"],"type":"string","description":"Bucket: silent_failure (tool returned ok but produced no useful output — empty list, wrong region, etc.), repeated_request (user had to ask for the same thing 2+ times because earlier turns didn't deliver), wrong_result (tool returned data but it answered a different question than the user asked), dissatisfaction (user is unhappy with a result and wants the team to know), missing_capability (user wants something the MCP can't do — 'why can't I…', 'I wish you could…'), other."},"severity":{"enum":["low","medium","high"],"type":"string","description":"Optional: low (minor papercut, user moved on), medium (user noticeably frustrated or had to repeat), high (user gave up / explicitly said this is broken)."},"tool_called":{"type":"string","pattern":"^leadbay_[a-z0-9_]{1,60}$","description":"Optional: the bare name of the registered Leadbay tool that disappointed, e.g. 'leadbay_pull_leads'. This is NOT a free-text field — any value that is not an actual registered tool name is dropped, so never encode context, detail, or user data here. Put context in the user-approved `message` instead."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","properties":{"_meta":{"type":"object","properties":{"region":{"type":"string"}}},"message":{"type":"string"},"reported":{"type":"boolean"},"_friction":{"type":"object","properties":{"message":{"type":"string"},"category":{"type":"string"},"severity":{"type":"string"},"tool_called":{"type":"string"}}}},"description":"Confirmation the report was sent. `reported: true` + a user-facing `message` the agent should show back to the user. The `_friction` block carries the analytics payload — the MCP server detects it and emits a `mcp friction reported` PostHog event containing only the fields the user approved."}},{"name":"leadbay_acknowledge_notification","annotations":{"title":"Acknowledge a Leadbay notification","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$8f","inputSchema":{"type":"object","required":["notification_id"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"archive":{"type":"boolean","description":"If true, archive the notification (won't appear in `archived=false` listings). If false / omitted, mark seen (resets firstSeenAt)."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."},"notification_id":{"type":"string","description":"UUID of the notification to acknowledge. Use the notification_id from `_meta.notifications[]` or `account_status.notifications[]`."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["acknowledged","notification_id","action"],"properties":{"action":{"enum":["seen","archive"],"type":"string"},"acknowledged":{"type":"boolean"},"notification_id":{"type":"string"}}}},{"name":"leadbay_get_artifact_runtime","annotations":{"title":"Artifact component kit","readOnlyHint":true,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$90","inputSchema":{"type":"object","properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"template":{"enum":["route_planner","triage_board"],"type":"string","description":"A finished board to publish as it is: \"route_planner\" (follow-ups on a map) or \"triage_board\" (today's leads as cards). Omit to get the runtime and usage guide for building a custom page."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["version"],"properties":{"html":{"type":"string","description":"The finished page. Publish it as it is."},"icon":{"type":"string","description":"The artifact's icon word."},"_meta":{"type":"object","description":"Server-injected notices, when present."},"title":{"type":"string","description":"The artifact's title (it is also the page's )."},"runtime":{"type":"string","description":"The self-contained JS the artifact inlines. Absent when a template is returned."},"version":{"type":"string","description":"Kit version; bump means the runtime changed."},"template":{"type":"string","description":"The template returned, when one was asked for."},"description":{"type":"string","description":"One sentence for the artifact's gallery card."},"usage_guide":{"type":"string","description":"Markdown guide for building the artifact. Absent when a template is returned."},"capabilities":{"type":"object","description":"The capability declaration to publish the page with."},"instructions":{"type":"string","description":"How to publish the template."}}}},{"name":"leadbay_get_basemap","annotations":{"title":"Country outline for a map board","readOnlyHint":true,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"**Do not call this tool.** The route planner template that\n`leadbay_get_artifact_runtime` serves calls it from inside the published page,\nto draw the grey country shape under its pins. It returns `{region, geojson}`\nfor the workspace's region (`fr` or `us`) — a large GeoJSON document that is\nuseless in a conversation.\n\nA board needs it because an artifact can only load its own published files,\nand an agent building a board for another user has no outline file to publish.\nWith this tool every board gets the same outline, for the right country.\n","inputSchema":{"type":"object","properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}},{"name":"leadbay_report_artifact_error","annotations":{"title":"Report an artifact runtime failure","readOnlyHint":false,"openWorldHint":true,"idempotentHint":false,"destructiveHint":false},"description":"$91","inputSchema":{"type":"object","required":["kind","surface"],"properties":{"code":{"type":"string","description":"Bounded error code (e.g. timeout, unavailable, QUOTA_EXCEEDED). NEVER a message — messages can carry user or API text."},"kind":{"enum":["bridge_unavailable","call_timeout","call_failed","parse_failed","options_empty","action_blocked","result_rejected"],"type":"string","description":"What went wrong. Exceptions: bridge_unavailable, call_timeout, call_failed, parse_failed. Outcomes (nothing threw): options_empty, action_blocked, result_rejected."},"tool":{"type":"string","description":"The Leadbay tool involved, when attributable to one."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"surface":{"enum":["call","field","action","resource","list"],"type":"string","description":"Which view-model surfaced it: call, field, action, resource, or list."},"kit_version":{"type":"string","description":"The @leadbay/components runtime version that emitted this."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}},{"name":"leadbay_find_new_leads","annotations":{"title":"Find new leads (net-new ICP search)","readOnlyHint":false,"openWorldHint":true,"idempotentHint":true,"destructiveHint":false},"description":"$92","inputSchema":{"type":"object","required":["count","request_id","_triggered_by"],"properties":{"lang":{"type":"string","description":"Output language (default: user's language)."},"count":{"type":"number","description":"Target DELIVERED leads, 1-50. With qualify:true this means n SURVIVORS of qualification, not n candidates examined."},"query":{"type":"string","description":"Natural-language ICP ask. Matches topic VOCABULARY — can surface vendors of a product as easily as buyers of it. Prefer example_lead. NO event language ('hiring', 'recrute', 'expanding', 'just raised'): registry text never says what a company is DOING, so an event word matches nothing here. Send the trigger to leadbay_set_qualification_questions or leadbay_refine_lead_targeting instead, and tell the user that is where it went."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"confirm":{"type":"boolean","description":"Explicit go-ahead, required only for a search that uses quota (qualify:true and/or channels). true = the user approved the quote, go ahead. false = a veto (returns mode:'needs_confirmation', uses nothing). Omitted on such a call → the tool withholds the submit and returns a free quote to show the user first. The default FREE search (no qualify, no channels) needs no confirm."},"dry_run":{"type":"boolean","description":"Validate + worst-case usage estimate + quota forecast. No job, uses nothing. Use before the first quota-using run of a session."},"filters":{"type":"object","properties":{"sectors":{"type":"array","items":{"type":"string"}},"locations":{"type":"array","items":{"type":"string"}},"employees_max":{"type":"number"},"employees_min":{"type":"number"}},"description":"HARD constraints (the seed only shapes ranking). Sector/location labels resolve at submit. `sectors` must be a label from Leadbay's own taxonomy (leadbay_list_sectors) — everyday wording like 'Professional Services' is not one; a spelling/plural slip is corrected for you, but a word with no label comes back as mode:'needs_sector_choice' with the labels to pick from. An unresolvable location is a 400 naming it.","additionalProperties":false},"novelty":{"enum":["org","none"],"type":"string","description":"org (default) = only companies NEW to the org (excludes org leads, lens members, CRM ids, prior MCP deliveries)."},"qualify":{"type":"boolean","description":"Run fresh AI qualification and drop candidates scoring below min_ai_score. Uses the org's usage quota for every candidate EXAMINED (survivor or not). Default false (free)."},"channels":{"type":"array","items":{"enum":["email","phone"],"type":"string"},"description":"Contact channels to find (email, phone). Uses the org's usage quota only when a value is found. Empty = free identity tier."},"max_cost":{"type":"number","description":"Usage cap for the whole job, in internal units. Leave it unset: the default (100000) covers any job. What the requested channels need is kept for them; a cap below that is refused, naming the minimum. Never show it to the user as money."},"request_id":{"type":"string","description":"REQUIRED idempotency key. Derive it from the ask (e.g. 'gyms-texas-2026-07-28'); REUSE the exact same value when retrying the same ask — a duplicate returns the SAME job instead of launching twice. Use a NEW value only for a genuinely new ask."},"title_gate":{"enum":["strict","prefer"],"type":"string","description":"strict = only leads with a matching known contact; prefer (default when contact_titles set) = matched first, rest flagged."},"example_lead":{"type":"object","properties":{"name":{"type":"string"},"location":{"type":"string"},"employees":{"type":"number"},"description":{"type":"string"}},"description":"A FICTIONAL typical ideal customer used as a look-alike seed — the highest-leverage input. Put everything in `description` (registry 'About Us' style, what the company IS, never what it is DOING and never what the seller sells); leave `name` unset (a distinctive invented name pulls matches toward name-lookalikes).","additionalProperties":false},"min_ai_score":{"type":"number","maximum":30,"minimum":-30,"description":"Disqualification floor on the [-30,+30] qualification DELTA (not the 0-100 fit score). Default 0. Lower to -30 to keep every evaluated lead with its evidence."},"wait_seconds":{"type":"number","description":"How long to poll before returning (default and maximum 45 — the whole call is bounded by it, so a slow submit shortens the poll; 0 = submit + one snapshot). Free searches usually finish inside the window; qualified exploration can take minutes — the result then carries still_running:true and the job_id to check with leadbay_lead_job_status."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"contact_titles":{"type":"array","items":{"type":"string"},"maxItems":10,"description":"Wanted decision-maker titles (max 10), matched semantically cross-language."},"exploration_cap":{"type":"number","description":"Max candidates the qualify gate may examine. Default min(3n,150), ceiling min(20n,1000)."},"exclude_lead_ids":{"type":"array","items":{"type":"string"},"description":"Caller-side novelty belt on top of the server-side one (max 500 ids — over that the call is refused, so drop the DELIVERED ids first: novelty:'org' already covers those, and the examined-but-rejected ones are what it misses)."}},"additionalProperties":false}},{"name":"leadbay_qualify_leads","annotations":{"title":"Qualify + get the right contact on known leads","readOnlyHint":false,"openWorldHint":true,"idempotentHint":false,"destructiveHint":false},"description":"$93","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"lang":{"type":"string","description":"Output language (default: user's language)."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"confirm":{"type":"boolean","description":"Explicit go-ahead for the work that uses quota (fresh qualification and/or channels). true = the user approved, go ahead. false = a veto (returns mode:'needs_confirmation', uses nothing). Omitted on such a call → the tool withholds the submit and returns a free quote to show the user first. A fully FREE call (qualify:false and no channels) needs no confirm."},"dry_run":{"type":"boolean","description":"Validate + worst-case usage + quota forecast. No job, uses nothing."},"qualify":{"type":"boolean","description":"Fresh AI qualification (default true; uses the org's usage quota for each lead needing fresh research+scoring, none when a fresh dossier exists). Owned disqualified leads come back WITH their negative evidence."},"channels":{"type":"array","items":{"enum":["email","phone"],"type":"string"},"description":"Channels to find (email, phone). Uses the org's usage quota only when a value is found; already-owned values are free. Empty = free identity tier."},"max_cost":{"type":"number","description":"Usage cap in internal units; leave it unset (default 100000). What the requested channels need is kept for them; a cap below that is refused, naming the minimum. Never show it as money."},"lead_refs":{"type":"array","items":{"type":"object","properties":{"name":{"type":"string"},"lead_id":{"type":"string","description":"Leadbay lead UUID."},"website":{"type":"string"},"location":{"type":"string","description":"Disambiguates name-only refs (city/region)."},"contact_id":{"type":"string","description":"Stable lead_contact id from a prior result — enrichment then targets EXACTLY this person, never a re-match."}},"additionalProperties":false},"description":"Companies to qualify (max 500). Each ref needs at least one identifying field. Duplicate lead_ids collapse into one item."},"request_id":{"type":"string","description":"Recommended idempotency key — REUSE the same value when retrying the same batch so a retry returns the SAME job instead of running it again."},"title_gate":{"enum":["strict","prefer"],"type":"string","description":"strict = only items with a matching known contact deliver a contact; prefer = matched first, rest flagged."},"wait_seconds":{"type":"number","description":"How long to poll before returning (default and maximum 45 — the whole call is bounded by it, so a slow submit shortens the poll; 0 = submit + one snapshot). Large or research-heavy batches can take minutes — the result then carries still_running:true and the job_id for leadbay_lead_job_status."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"contact_titles":{"type":"array","items":{"type":"string"},"description":"Wanted decision-maker titles (max 10), matched semantically."},"prior_deliveries":{"type":"object","properties":{"limit":{"type":"number"},"since":{"type":"string","description":"ISO instant lower bound."},"job_id":{"type":"string"}},"description":"Selector expanding the org's past MCP deliveries into refs — delivered leads stay re-readable after result expiry. Combine with lead_refs or use alone.","additionalProperties":false}},"additionalProperties":false}},{"name":"leadbay_bulk_qualify_leads","annotations":{"title":"Bulk-qualify next N leads","readOnlyHint":false,"openWorldHint":true,"idempotentHint":true,"destructiveHint":false},"description":"$94","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"count":{"type":"number","description":"How many fresh leads to qualify (default 10, max 25)"},"lensId":{"type":"number","description":"Lens id (escape hatch — defaults to active)"},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"leadIds":{"type":"array","items":{"type":"string"},"description":"Explicit lead UUIDs to qualify (skips the auto-pagination)"},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"total_budget_ms":{"type":"number","description":"Total polling budget in ms (default 300000)"},"per_lead_budget_ms":{"type":"number","description":"Polling budget per lead in ms (default 90000)"},"wait_for_completion":{"type":"boolean","description":"When false, launch qualification and return `{status:'running', notification_id, lead_ids, lens_id}` immediately. Poll leadbay_qualify_status with them. Default is true for 0.6.x backwards compatibility."}},"additionalProperties":false},"outputSchema":{"type":"object","anyOf":[{"required":["qualified","still_running","failed","quota_exceeded"]},{"required":["status","notification_id","lead_ids","launched_count","failed","quota_exceeded","lens_id","_meta"]}],"required":["failed","quota_exceeded"],"properties":{"_meta":{"type":"object","properties":{"region":{"type":"string"}},"description":"Operator context: region."},"failed":{"type":"array","items":{"type":"object"},"description":"Leads whose web_fetch launch failed (per-lead error)."},"status":{"type":"string","description":"`running` when wait_for_completion=false; absent on the legacy blocking result."},"lens_id":{"type":"number","description":"The lens id the qualification ran against. Present on every successful return."},"message":{"type":"string","description":"Human-readable summary; absent on the happy path."},"lead_ids":{"type":"array","items":{"type":"string"}},"exhausted":{"type":"boolean","description":"True if the lens's wishlist had no more unqualified leads to qualify."},"qualified":{"type":"array","items":{"type":"object"},"description":"Leads whose qualification finished within budget. Each entry: lead_id, qualification_summary{answered,total,avg_qualification_boost}, signals_count."},"still_running":{"type":"array","items":{"type":"object"},"description":"Leads launched but whose qualification did not complete within budget. Re-poll via leadbay_qualify_status with notification_id (when present) or lead_ids + lens_id."},"launched_count":{"type":"number"},"quota_exceeded":{"type":"boolean","description":"True if 429 was hit mid-fanout. Already-launched leads keep polling; further launches stopped."},"notification_id":{"type":["string","null"],"description":"The backend's job id. Carry it to leadbay_qualify_status."},"total_unqualified_found":{"type":"number"}}}},{"name":"leadbay_enrich_titles","annotations":{"title":"Enrich contact titles across leads","readOnlyHint":false,"openWorldHint":true,"idempotentHint":true,"destructiveHint":false},"description":"$95","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"email":{"type":"boolean","description":"Enrich emails (default true)"},"phone":{"type":"boolean","description":"Enrich phone numbers (default false)"},"lensId":{"type":"number","description":"Lens id (escape hatch — defaults to active)"},"titles":{"type":"array","items":{"type":"string"},"description":"Job titles to enrich. Omit to discover what's available without launching."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"confirm":{"type":"boolean","description":"Explicit go-ahead for the email/phone reveal (uses quota). true = go ahead and launch. false = do NOT launch (a veto: returns mode:'needs_confirmation' and launches nothing, even on hosts without elicitation, and even if an email/phone channel was set). Omitted (and no explicit email/phone channel) → an elicitation-capable host asks the user before launching; a decline returns mode:'needs_confirmation'. Passing email:true/phone:true also counts as consent."},"dry_run":{"type":"boolean","description":"If true, don't launch — only preview."},"leadIds":{"type":"array","items":{"type":"string"},"description":"Lead UUIDs to enrich. Omit to use the top page of the active lens's wishlist."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"candidateCount":{"type":"number","description":"When leadIds is omitted, how many top-of-wishlist leads to use (default 25)"}},"additionalProperties":false},"outputSchema":{"type":"object","properties":{"mode":{"type":"string","description":"'discover' | 'preview_only' | 'dry_run' | 'needs_confirmation' | 'already_launched' | 'launch_in_flight' | 'launched'."},"email":{"type":"boolean"},"phone":{"type":"boolean"},"reused":{"type":"boolean","description":"True when an identical launch inside the 5-minute window was reused instead of spending quota again."},"status":{"type":"string","description":"'quota_exceeded' on 429. Otherwise mode is set instead."},"titles":{"type":"array","items":{"type":"string"},"description":"Titles ordered (echoed at launch)."},"message":{"type":"string","description":"Operator-facing summary of what happened."},"preview":{"type":"object","description":"Backend BulkEnrichPreview payload (preview_only/dry_run/launched modes)."},"launched":{"type":"boolean","description":"True when an enrichment job is in flight on the backend."},"lead_ids":{"type":"array","items":{"type":"string"},"description":"The leads this run enriched. Carry them to leadbay_bulk_enrich_status for per-lead progress; they also work when the notification is archived."},"launched_at":{"type":"string","description":"ISO timestamp of the (re-used or fresh) launch."},"next_action":{"type":"string","description":"Concrete next-step instruction for the agent."},"would_launch":{"type":"object","description":"What dry_run WOULD have launched (titles, email, phone)."},"auto_included":{"type":"array","items":{"type":"string"},"description":"Backend's auto_included_titles (discover mode)."},"notification_id":{"type":["string","null"],"description":"The backend's job id. Carry it to leadbay_bulk_enrich_status. Null when the backend returned none — then poll by lead_ids instead."},"recommendations":{"type":"array","items":{"type":"string"},"description":"Backend's title_suggestions (discover mode)."},"available_titles":{"type":"array","items":{"type":"string"},"description":"Titles available across the selection (discover/preview_only modes)."},"credits_remaining":{"type":["number","string","null"],"description":"Advisory-only balance (billing.ai_credits), present in discover / preview_only / dry_run modes. Enrichment is gated by QUOTA (email + phone reveals consume the per-window allowance in leadbay_account_status), NOT by this number — do NOT present it as a gate, do NOT say 'you have N credits', and NEVER refuse enrichment because it's low or zero. Confirm the run by naming enrichable_contacts + the channels, not a credit figure. Null = billing unavailable. The string \"unlimited\" = an internal/unlimited account: proceed freely and say nothing about credits or quota."},"enrichable_contacts":{"type":"number","description":"Count of enrichable contacts at preview time."},"previously_enriched":{"type":"array","items":{"type":"string"},"description":"Titles previously enriched on this selection (discover mode)."},"retry_after_seconds":{"type":["number","null"],"description":"Seconds until quota resets (quota_exceeded status)."},"selected_lead_count":{"type":"number","description":"How many leads the selection covers."},"already_enriched_contacts":{"type":"number","description":"preview_only mode: contacts with these titles on the selected leads that your organization already enriched. Above 0 means the work is done, not that nobody has these titles."},"seconds_since_original_launch":{"type":"number","description":"Age of the re-used bulk record (already_launched mode)."}},"description":"Branchy return shape; the `mode` (or `status`) field tells the agent which branch it got. Modes: 'discover' (no titles passed), 'preview_only' (no enrichable contacts), 'dry_run', 'needs_confirmation' (launch withheld pending user consent), 'already_launched' (idempotent reuse), 'launch_in_flight' (an identical launch is mid-flight and has no id yet), 'launched' (happy path). Status: 'quota_exceeded' (429)."}},{"name":"leadbay_adjust_audience","annotations":{"title":"Adjust lens audience filters","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":true},"description":"$96","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"sizes":{"type":"array","items":{"type":"object","properties":{"max":{"type":"number"},"min":{"type":"number"}}},"description":"Company size buckets, e.g. [{min:30,max:300}]"},"lensId":{"type":"number","description":"Lens id (escape hatch)"},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"sectors":{"type":"array","items":{"type":"string"},"description":"Sector free-text (e.g. ['Healthcare', 'Engineering']) or ids — auto-resolved"},"lensName":{"type":"string","description":"Target a lens BY NAME (e.g. 'Joinery') instead of the active one. Resolved against your lenses — edit-only, does NOT switch your active lens. Unknown/ambiguous names are surfaced to pick from. Takes effect only when lensId is not given."},"locations":{"type":"array","items":{"type":"string"},"description":"Geographic scope — free text (e.g. ['Indre-et-Loire', 'Texas', 'Austin']) or admin-area ids. Resolved via /geo/search at any level from state down to city (state / région / département / county / city). NEVER a country name — this workspace serves exactly ONE country, so a whole-country ask means passing NO location at all (rejected with COUNTRY_LEVEL_LOCATION). Place names go HERE, never in sectors/keywords."},"sector_ids":{"type":"array","items":{"type":"string"},"description":"Explicit sector ids (skips taxonomy lookup)"},"newLensName":{"type":"string","description":"Name to use when this composite has to clone the default lens (otherwise auto-named)"},"location_ids":{"type":"array","items":{"type":"string"},"description":"Explicit admin-area ids (skips /geo/search resolution). Sub-country areas only — a country name here is rejected with COUNTRY_LEVEL_LOCATION."},"save_for_org":{"type":"boolean","description":"Admin only — propagate the change to the org-level lens for everyone (default false: per-user draft)"},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"exclude_sectors":{"type":"array","items":{"type":"string"},"description":"Sectors to exclude (free text or ids)"},"exclude_locations":{"type":"array","items":{"type":"string"},"description":"Locations to exclude (free text or ids). Sub-country areas only — excluding a country is meaningless on a single-country workspace and is rejected."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["status"],"properties":{"_meta":{"type":"object"},"lenses":{"type":"array","items":{"type":"object"},"description":"On 'lens_not_found': the user's lenses [{id, name}] to pick from."},"status":{"type":"string","description":"'applied', 'ambiguous_sectors', 'ambiguous_locations', 'country_level_location', 'lens_not_found', or 'ambiguous_lens'."},"matches":{"type":"array","items":{"type":"object"},"description":"On 'ambiguous_lens': the lenses [{id, name}] the name matched."},"message":{"type":"string"},"lens_used":{"type":"object","description":"Resolved lens metadata: {id, name, was_draft, was_new, active_lens_changed, save_for_org}."},"lens_query":{"type":"string","description":"On 'lens_not_found' / 'ambiguous_lens': the lensName the user asked for."},"filter_applied":{"type":"object","description":"The merged FilterPayload that was POSTed to the lens."},"country_locations":{"type":"array","items":{"type":"object"},"description":"On 'country_level_location': per offending value {value, param, kind, country, axis, kept}. A country name is never a location criterion — each workspace serves exactly ONE country. The lens was NOT modified. The recovery BRANCHES on `country_locations[].axis` and `[].kind`; `hint` states the one for THIS call — follow it verbatim. When the country was the ONLY scope, or on ANY non-foreign `exclude`, the answer is to write NOTHING at all — re-calling with the value merely dropped persists a scope that inverts the request. Never retry with another spelling or a nearby city."},"sector_ambiguities":{"type":"array","items":{"type":"object"},"description":"Per ambiguous text: {sector_text, matches:[{id, name, score}]}. Agent picks an id and re-calls."},"location_ambiguities":{"type":"array","items":{"type":"object"},"description":"On 'ambiguous_locations': per text {location_text, matches:[{id, name, country, level, score}]}. Agent picks an id and re-calls via the SAME axis the text came from — an include text → location_ids; a text from exclude_locations → exclude_locations (NOT location_ids, which would include the area the user asked to exclude). The `message` field names the correct param per text."}},"description":"Return shapes: 'applied' on success; 'ambiguous_sectors' when free-text sectors matched multiple candidates (re-call with sector_ids); 'ambiguous_locations' when free-text locations didn't resolve to one area — re-call with the chosen id via the SAME axis it came from (an include pick → location_ids; an EXCLUDE pick → exclude_locations, NOT location_ids, which would include it); 'country_level_location' when a country-level value was passed as a location (nothing was read or written; read `hint` — re-calling without the value is often itself wrong); 'lens_not_found' / 'ambiguous_lens' when a lensName didn't resolve to exactly one lens (re-call with lensId or an exact lensName)."}},{"name":"leadbay_refine_lead_targeting","annotations":{"title":"Refine the audience prompt","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":true},"description":"$97","inputSchema":{"type":"object","required":["prompt","_triggered_by"],"properties":{"prompt":{"type":"string","description":"Refinement instruction (free text)"},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"dry_run":{"type":"boolean","description":"If true, return the call shape WITHOUT setting the prompt"},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"clarification_poll_gap_ms":{"type":"number","description":"Gap between polls in ms (default 5000)"},"clarification_poll_attempts":{"type":"number","description":"How many times to poll for a clarification after setting (default 2)"}},"additionalProperties":false},"outputSchema":{"type":"object","properties":{"_meta":{"type":"object"},"status":{"type":"string","description":"'applied' (prompt set; intelligence regenerating) or 'clarification_pending' (telephone path)."},"dry_run":{"type":"boolean","description":"True when dry_run:true was passed (no state change)."},"message":{"type":"string","description":"Operator-facing summary."},"would_call":{"type":"object","description":"Dry-run preview of the POST that would have been issued."},"next_action":{"type":"string","description":"Concrete next-step instruction for the agent."},"clarification":{"type":"object","description":"ClarificationPayload returned by the backend (clarification_pending path)."},"clarified_via_elicit":{"type":"boolean","description":"True when the clarification was answered via the client's elicitation UI (not via telephone)."},"computing_intelligence":{"type":"boolean","description":"True when intelligence is regenerating after the prompt set."}},"description":"Multiple return shapes by status. dry_run, applied (with optional clarified_via_elicit), or clarification_pending."}},{"name":"leadbay_answer_clarification","annotations":{"title":"Answer pending clarification","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":true},"description":"$98","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"option_id":{"type":"string","description":"Id of one of the clarification's options"},"text_answer":{"type":"string","description":"Free-text answer (overrides option_id)"},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["status"],"properties":{"hint":{"type":"string","description":"Operator-facing next-step (no_pending_clarification path)."},"_meta":{"type":"object"},"status":{"type":"string","description":"'answered' (recorded; intelligence regenerating) or 'no_pending_clarification' (nothing to answer)."},"message":{"type":"string"},"recorded_as_user_prompt":{"type":"boolean","description":"True when the answer was stored as the org's new user_prompt."}}}},{"name":"leadbay_report_outreach","annotations":{"title":"Report outreach to Leadbay","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":false},"description":"$99","inputSchema":{"type":"object","required":["note","verification","_triggered_by"],"properties":{"note":{"type":"string","description":"1-2 sentence summary of what was done (e.g. 'Sent intro email to CTO citing Hornsea 3 contract')"},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"dry_run":{"type":"boolean","description":"If true, return what WOULD be called without writing anything"},"lead_id":{"type":"string","description":"Single lead UUID (use lead_ids for bulk)"},"lead_ids":{"type":"array","items":{"type":"string"},"description":"Bulk: many lead UUIDs (epilogue applies to all; notes fan out)"},"contact_id":{"type":"string","description":"Optional, with lead_id only: the org contact id of the person the outreach went to. The note and the epilogue status are then written on that person, and still show on the lead's Monitor row."},"verification":{"type":"object","required":["source","ref"],"properties":{"ref":{"type":"string"},"source":{"type":"string"}},"description":"REQUIRED. Proof the action actually happened. source: gmail_message_id|calendar_event_id|user_confirmed. ref: the message id, event id, or the user's confirming text.","additionalProperties":false},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"epilogue_status":{"type":"string","description":"Optional: STILL_CHASING | COULD_NOT_REACH_STILL_TRYING | INTEREST_VALIDATED_OR_MEETING_PLANED | NOT_INTERESTED_LOST"}},"additionalProperties":false},"outputSchema":{"type":"object","properties":{"_meta":{"type":"object","properties":{"region":{"type":"string"}},"description":"Operator context: region."},"notes":{"type":"object","properties":{"failed":{"type":"array","items":{"type":"object"}},"succeeded":{"type":"array","items":{"type":"object"}}},"description":"Per-lead note-write outcome (split into succeeded / failed sub-arrays)."},"dry_run":{"type":"boolean"},"epilogue":{"type":"object","properties":{"error":{"type":"string"},"status":{"type":["string","null"]},"applied":{"type":"boolean"}},"description":"Epilogue status outcome: status (the wire-format string written to /leads/epilogue, or null when not requested), applied (true/false), error (when applied=false)."},"contact_id":{"type":"string","description":"Present when the note and the status were written on this person rather than on the lead."},"verification":{"type":"object","properties":{"ref":{"type":"string"},"source":{"type":"string"}},"description":"Effective verification used (after elicit override). Useful for the client UI to render \"logged with proof X\". When source=user_confirmed AND ctx.elicit was available, ref is the user's literal text typed into the client; otherwise it's the agent-supplied ref."},"confirmed_via":{"type":"string","description":"Audit trail of how verification was obtained: 'elicit' (user typed into client UI — anti-poisoning), 'agent_supplied' (legacy path; user_confirmed source with no elicit), 'non_user_confirmed' (gmail_message_id or calendar_event_id — agent can't fabricate these)."},"already_logged":{"type":"array","items":{"type":"string"},"description":"Lead ids skipped because a note with this exact gmail_message_id or calendar_event_id is already there. Nothing was written for them."},"would_write_notes":{"type":"array","items":{"type":"object"},"description":"On dry_run: the per-lead POST shapes that WOULD be issued."},"would_set_epilogue":{"type":["object","null"],"description":"On dry_run: the epilogue POST shape that WOULD be issued."}},"description":"Either the dry_run shape (dry_run:true with would_write_notes / would_set_epilogue) OR the live result (notes:{succeeded,failed} + epilogue:{status,applied,error?} + verification + _meta). Schema declares both shapes; the dry_run discriminator picks which sub-shape applies."}},{"name":"leadbay_set_prospecting_action","annotations":{"title":"Set or clear a prospecting action","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":false},"description":"$9a","inputSchema":{"type":"object","required":["lead_id","action","selected","_triggered_by"],"properties":{"action":{"enum":["STILL_CHASING","COULD_NOT_REACH_STILL_TRYING","INTEREST_VALIDATED_OR_MEETING_PLANED","NOT_INTERESTED_LOST"],"type":"string","description":"The prospecting action. The EPILOGUE_-prefixed form is accepted too."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"lead_id":{"type":"string","description":"Lead UUID."},"selected":{"type":"boolean","description":"true to turn the action on for today, false to turn it off."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["lead_id","action","selected"],"properties":{"action":{"type":"string","description":"The short form, e.g. STILL_CHASING."},"lead_id":{"type":"string"},"selected":{"type":"boolean","description":"The state the action is now in."}}}},{"name":"leadbay_send_feedback","annotations":{"title":"Send feedback to the Leadbay team","readOnlyHint":false,"openWorldHint":true,"idempotentHint":false,"destructiveHint":false},"description":"$9b","inputSchema":{"type":"object","required":["message"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"message":{"type":"string","description":"The user's feedback, in their own words. Confirm the wording with the user BEFORE calling — this is sent to the Leadbay team. Cap 4000 chars."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."},"associated_error_id":{"type":"string","description":"Optional: a Sentry event id to attach this feedback to (e.g. the id from an error the user just hit), so the team sees the feedback on that exact issue."}},"additionalProperties":false},"outputSchema":{"type":"object","properties":{"sent":{"type":"boolean"},"_meta":{"type":"object","properties":{"region":{"type":"string"}}},"message":{"type":"string"}},"description":"Whether the feedback reached the Leadbay team. `sent: true` means it landed in the team's inbox."}},{"name":"leadbay_import_leads","annotations":{"title":"Import leads from list/file","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$9c","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"domains":{"type":"array","items":{"type":"object","required":["domain"],"properties":{"name":{"type":"string","description":"Optional display name override; defaults to the domain."},"domain":{"type":"string","description":"Company domain (e.g. 'apple.com'). Protocol/path are stripped."}},"additionalProperties":false},"description":"Mode A: list of company domains to map to Leadbay leadIds. Mutually exclusive with `records`."},"dry_run":{"type":"boolean","description":"If true, run preprocess only — do NOT commit lead-CRM linking. Note: an import row still appears in the user's CRM-imports list as 'incomplete'. Use to verify input format / wizard reachability without polluting the CRM."},"records":{"type":"array","items":{"type":"object"},"description":"Mode B: arbitrary CSV-shaped rows. Each record is an object whose keys are column names and values are scalar (string/number/boolean/null). Mutually exclusive with `domains`. Must be accompanied by `mappings.fields`. The tool synthesizes a CSV from the union of all record keys."},"mappings":{"type":"object","required":["fields"],"properties":{"fields":{"type":"object","description":"$9d"},"statuses":{"type":"object","description":"Optional map of raw CSV status-cell text → lead status (rarely needed). Keys are the verbatim cell strings; values must be one of DEFAULT, INBOUND, UNWANTED, WANTED, LOST, WON (case-insensitive). Defaults to {}."},"custom_fields":{"type":"object","description":"Ergonomic shorthand: `{CsvColumn: <number-id>}` or `{CsvColumn: '<field-name>'}` for custom-field mappings. Resolved against the org's /crm/custom_fields catalog before the import is committed. Mutually exclusive with `fields[col] = 'CUSTOM.<id>'` for the same column. Useful when the agent doesn't want to deal with the 'CUSTOM.<id>' wire format."},"default_status":{"enum":["DEFAULT","INBOUND","UNWANTED","WANTED","LOST","WON",null],"type":["string","null"],"description":"Optional default lead status applied to rows without an explicit status. One of DEFAULT, INBOUND, UNWANTED, WANTED, LOST, WON (case-insensitive). Defaults to null."}},"description":"Mode B: how each CSV column maps to Leadbay's CRM field schema. Required when `records` is supplied; ignored otherwise."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"total_budget_ms":{"type":"number","description":"Overall cap across all phases (default 300000ms)."},"per_phase_budget_ms":{"type":"number","description":"Single poll-loop cap (default 60000ms)."},"wait_for_completion":{"type":"boolean","description":"When false, upload the rows and return `{status:'running', importIds}` immediately. Poll leadbay_import_status({importIds, dry_run}) — those importIds are the backend's own and keep working from any conversation. Default is true."}},"additionalProperties":false},"outputSchema":{"type":"object","anyOf":[{"required":["leads","not_imported","importIds","region","_meta"]},{"required":["status","importIds","progress","region","_meta"]}],"required":["importIds","region","_meta"],"properties":{"_meta":{"type":"object"},"leads":{"type":"array","items":{"type":"object"},"description":"Imported leads. Domains mode: [{domain, leadId, name}]. Records mode: [{rowId, domain?, leadId, name}]."},"region":{"type":"string"},"status":{"type":"string","description":"`running` when wait_for_completion=false; absent on the legacy blocking result."},"dry_run":{"type":"boolean","description":"True when dry_run:true was passed (preprocess only, no CRM commit)."},"row_ids":{"type":"array","items":{"type":"string"},"description":"Records mode only: the synthetic MCP_ROW_ID of each input row, in the order you passed `records[]`. leadbay_import_status reports recovered leads by that id — use this to map them back to your source rows."},"progress":{"type":"object","description":"Current async import progress when wait_for_completion=false."},"cancelled":{"type":"boolean","description":"True when the HOST cancelled the call (a user Cancel, or the host's own request timeout); a Leadbay-side or phase-budget timeout reports `timed_out` instead. Rows already uploaded keep importing on Leadbay — poll leadbay_import_status with importIds, but a chunk cancelled before its mappings were committed can read `running` without ever progressing; if the counts stop moving, say so rather than polling on. Rows past the interruption may never have been uploaded; re-run the import for those rows only."},"importIds":{"type":"array","items":{"type":"string"},"description":"Backend file-import handles (one per chunk of ≤100 rows)."},"timed_out":{"type":"boolean","description":"True when a blocking call ran out of poll budget. The import is still running server-side — poll leadbay_import_status(importIds). Do NOT re-issue the import."},"not_imported":{"type":"array","items":{"type":"object"},"description":"Inputs that did NOT yield a leadId. Each entry has a `reason` ('malformed', 'NO_MATCH', 'TIMEOUT', etc.) plus the input echo."},"rows_pending_upload":{"type":"number","description":"Rows from later chunks that were never uploaded before the budget ran out. These are NOT running anywhere; re-import just those rows."}}}},{"name":"leadbay_import_and_qualify","annotations":{"title":"Import + qualify leads","readOnlyHint":false,"openWorldHint":true,"idempotentHint":true,"destructiveHint":false},"description":"$9e","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"lensId":{"type":"number","description":"Lens id (escape hatch — defaults to active)."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"domains":{"type":"array","items":{"type":"object","required":["domain"],"properties":{"name":{"type":"string","description":"Optional display name override; defaults to the domain."},"domain":{"type":"string","description":"Company domain (e.g. 'apple.com'). Protocol/path are stripped."}},"additionalProperties":false},"description":"Mode A: list of company domains to map to Leadbay leadIds. Mutually exclusive with `records`."},"dry_run":{"enum":[true,false,"preview"],"type":["boolean","string"],"description":"Optional. `true` runs preprocess only (no commit, no qualify). `'preview'` runs preprocess and returns the wizard's per-column AI mapping hints + sample rows + custom-field candidates from the org catalog so the agent can choose a mapping. Default: false (full flow)."},"records":{"type":"array","items":{"type":"object"},"description":"Mode B: arbitrary CSV-shaped rows. Each record is an object whose keys are column names and values are scalar (string/number/boolean/null). Mutually exclusive with `domains`. Accompany with `mappings` UNLESS using `dry_run: 'preview'` — preview returns the wizard's mapping suggestions without requiring a mapping up front."},"mappings":{"type":"object","properties":{"fields":{"type":"object","description":"Object whose keys are CSV column names and whose values are either StandardCrmFieldType (LEAD_NAME, LEAD_WEBSITE, ..., CONTACT_TITLE) or 'CUSTOM.<id>'. Discover via leadbay_list_mappable_fields. At least one entry must target LEADBAY_ID, CRM_ID, SIREN, LEAD_NAME, or LEAD_WEBSITE. Use leadbay_resolve_import_rows to prepare LEADBAY_ID values from messy user files. Contact exports and embedded owner/contact lists should map CONTACT_EMAIL/PHONE/TITLE/name fields while preserving parent lead identity; expand structured people into repeated parent rows. HubSpot/source links should map to CUSTOM.<id> fields created or discovered before import."},"statuses":{"type":"object","description":"Optional map of raw CSV status-cell text → lead status. Values must be one of DEFAULT, INBOUND, UNWANTED, WANTED, LOST, WON (case-insensitive); keys are the verbatim cell strings."},"custom_fields":{"type":"object","description":"Ergonomic shorthand: `{CsvColumn: <number-id>}` or `{CsvColumn: '<field-name>'}` for custom-field mappings. Resolved against /crm/custom_fields catalog."},"default_status":{"enum":["DEFAULT","INBOUND","UNWANTED","WANTED","LOST","WON",null],"type":["string","null"],"description":"Optional default lead status. One of DEFAULT, INBOUND, UNWANTED, WANTED, LOST, WON (case-insensitive)."}},"description":"How each CSV column maps to Leadbay's CRM field schema. Required for records-mode WITHOUT `dry_run: 'preview'`; ignored otherwise. The `fields` sub-property is REQUIRED only when at least one column maps to a StandardCrmFieldType target; pure custom-field mappings can be supplied via `custom_fields` shorthand alone.","additionalProperties":false},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"total_budget_ms":{"type":"number","description":"Total wall-clock budget across import + qualify in ms (default 900000). When exhausted, the response returns lead_ids + lens_id for resume via leadbay_qualify_status."},"per_lead_budget_ms":{"type":"number","description":"Polling budget per lead in ms (default 90000)."},"per_phase_budget_ms":{"type":"number","description":"Per-phase budget for the import wizard (default 300000); mirrors leadbay_import_leads."},"wait_for_completion":{"type":"boolean","description":"When false, upload the rows and return `{kind:'result', status:'running', import_ids}` immediately. Poll leadbay_import_status({importIds, dry_run}); the qualify phase does NOT run on this path — call leadbay_bulk_qualify_leads on the imported leads yourself. Default is true."},"skip_already_qualified":{"type":"boolean","description":"When true (default), skips web_fetch launch on leads whose ai_agent_lead_score is already non-null. Saves quota. Set false to force fresh re-qualification."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["kind","region","_meta"],"properties":{"kind":{"type":"string","description":"'result' (full flow) or 'preview' (dry_run='preview' mapping diagnostics)."},"_meta":{"type":"object"},"notes":{"type":"array","items":{"type":"string"},"description":"Operator notes (e.g., catalog fetch errors)."},"failed":{"type":"array","items":{"type":"object"},"description":"Per-lead errors observed during qualification."},"region":{"type":"string"},"reused":{"type":"boolean","description":"True when an identical launch was reused within the idempotency window."},"status":{"type":"string","description":"`running` when wait_for_completion=false."},"dry_run":{"type":"boolean","description":"True when dry_run:true was passed."},"lens_id":{"type":"number","description":"Lens the qualification ran against."},"row_ids":{"type":"array","items":{"type":"string"},"description":"Records mode only: the synthetic row id of each input row, in the order you passed `records[]`. leadbay_import_status reports recovered leads by that id — use this to map them back to your source rows."},"imported":{"type":"array","items":{"type":"object"},"description":"Leads that landed in CRM. Each: {leadId, domain?, name, rowId?}."},"lead_ids":{"type":"array","items":{"type":"string"},"description":"Leads the qualify phase covers. Pass to leadbay_qualify_status for per-lead detail."},"cancelled":{"type":"boolean","description":"True when the HOST cancelled the call (a user Cancel, or the host's own request timeout); a budget timeout reports `budget_exhausted` instead. The import and the qualifications already launched keep running on Leadbay, except a chunk cancelled before its mappings were committed, which can read `running` without ever progressing — if the counts stop moving, say so rather than polling on. Poll leadbay_import_status with the `import_ids` values passed as `importIds`, and leadbay_qualify_status with lead_ids + lens_id. `still_running` can also hold leads whose qualification was never launched; those need a fresh qualification, which qualify_status cannot start."},"import_id":{"type":"string","description":"Backend file-import handle (preview shape)."},"qualified":{"type":"array","items":{"type":"object"},"description":"Leads whose qualification settled within budgets."},"timed_out":{"type":"boolean","description":"True when the underlying import ran out of poll budget. Still running server-side — poll leadbay_import_status(import_ids). Do NOT re-issue the import."},"import_ids":{"type":"array","items":{"type":"string"},"description":"Backend file-import handles (one per chunk)."},"not_in_lens":{"type":"array","items":{"type":"string"},"description":"Lead ids that aren't members of the active lens — backend won't qualify them."},"sample_rows":{"type":"array","items":{"type":"object"},"description":"First few rows of the preprocessed sample (preview shape)."},"not_imported":{"type":"array","items":{"type":"object"},"description":"Inputs that didn't land. Each has a `reason` plus the input echo."},"mapping_hints":{"type":"array","items":{"type":"object"},"description":"Per-column AI-confidence suggestions (preview shape)."},"quota_blocked":{"type":"boolean","description":"True when quota was exhausted before launching all leads."},"still_running":{"type":"array","items":{"type":"object"},"description":"Leads still being qualified at deadline; agent calls leadbay_qualify_status with lead_ids + lens_id."},"chosen_budgets":{"type":"object","description":"Adaptive budgets the composite selected (when caller didn't override): {per_lead_budget_ms, total_budget_ms, per_phase_budget_ms, wall_clock_estimate_ms, strategy}."},"quota_exceeded":{"type":"boolean"},"budget_exhausted":{"type":"boolean","description":"True when total_budget_ms hit before all leads finished."},"rows_pending_upload":{"type":"number","description":"Rows from later chunks that never reached the backend. These are NOT running anywhere; re-import just those rows."},"seconds_since_original":{"type":"number"},"custom_field_candidates":{"type":"array","items":{"type":"object"},"description":"Org custom fields that match unmapped columns (preview shape)."},"skipped_already_qualified":{"type":"array","items":{"type":"string"},"description":"Lead ids skipped because ai_agent_lead_score was already non-null (skip_already_qualified=true)."}},"description":"Two return shapes: kind:'preview' (when dry_run='preview') with mapping hints; kind:'result' (default) with imported + qualified leads + lead_ids/lens_id to resume via leadbay_qualify_status."}},{"name":"leadbay_add_contact","annotations":{"title":"Add a contact","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":false},"description":"$9f","inputSchema":{"type":"object","required":["lead_id","first_name","last_name"],"properties":{"email":{"type":"string","description":"Contact email (optional)."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"lead_id":{"type":"string","description":"UUID of the parent company (lead) to attach the contact to. The contact is created on this company."},"job_title":{"type":"string","description":"Contact job title (optional)."},"last_name":{"type":"string","description":"Contact last name."},"first_name":{"type":"string","description":"Contact first name."},"phone_number":{"type":"string","description":"Contact phone number (optional, free-form)."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."},"linkedin_page":{"type":"string","description":"Contact LinkedIn profile URL (optional)."}},"additionalProperties":false}},{"name":"leadbay_remove_contact","annotations":{"title":"Remove a contact","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":true},"description":"$a0","inputSchema":{"type":"object","required":["contact_id"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"contact_id":{"type":"string","description":"UUID of the contact to remove (the contact's own `id`, e.g. from leadbay_research_lead_by_id's contacts list — NOT the parent lead id)."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}},{"name":"leadbay_pin_contact","annotations":{"title":"Pin a contact","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$a1","inputSchema":{"type":"object","required":["contact_id"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"contact_id":{"type":"string","description":"UUID of the contact to pin (the contact's own `id` — NOT the parent lead id)."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}},{"name":"leadbay_unpin_contact","annotations":{"title":"Unpin a contact","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$a2","inputSchema":{"type":"object","required":["contact_id"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"contact_id":{"type":"string","description":"UUID of the contact to unpin (the contact's own `id` — NOT the parent lead id)."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}},{"name":"leadbay_update_contact","annotations":{"title":"Update a contact","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":true},"description":"$a3","inputSchema":{"type":"object","required":["contact_id","first_name","last_name"],"properties":{"email":{"type":["string","null"],"description":"Contact email. Omit it to leave it unchanged. Pass null to ERASE it — an erase rewrites the whole contact, so that call must also carry every other optional field, or it is refused."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"job_title":{"type":["string","null"],"description":"Contact job title. Omit it to leave it unchanged. Pass null to ERASE it — an erase rewrites the whole contact, so that call must also carry every other optional field, or it is refused."},"last_name":{"type":"string","description":"Contact last name — REQUIRED even on an edit. Pass the current value if you're not changing it."},"contact_id":{"type":"string","description":"UUID of the contact to edit (the contact's own `id` — NOT the parent lead id)."},"first_name":{"type":"string","description":"Contact first name — REQUIRED even on an edit. Pass the current value if you're not changing it."},"phone_number":{"type":["string","null"],"description":"Contact phone (free-form). Omit it to leave it unchanged. Pass null to ERASE it — an erase rewrites the whole contact, so that call must also carry every other optional field, or it is refused."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."},"linkedin_page":{"type":["string","null"],"description":"Contact LinkedIn URL. Omit it to leave it unchanged. Pass null to ERASE it — an erase rewrites the whole contact, so that call must also carry every other optional field, or it is refused."}},"additionalProperties":false}},{"name":"leadbay_enrich_contacts","annotations":{"title":"Enrich contacts for a lead","readOnlyHint":false,"openWorldHint":true,"idempotentHint":false,"destructiveHint":true},"description":"$a4","inputSchema":{"type":"object","required":["leadId","contactId"],"properties":{"email":{"type":"boolean","description":"Enrich email address (default: true)"},"phone":{"type":"boolean","description":"Enrich phone number (default: true)"},"leadId":{"type":"string","description":"Lead UUID (required)"},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"contactId":{"type":"string","description":"Contact UUID (required)"},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}},{"name":"leadbay_create_custom_field","annotations":{"title":"Create CRM custom field","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$a5","inputSchema":{"type":"object","required":["name","_triggered_by"],"properties":{"name":{"type":"string","description":"User-visible custom field name, e.g. 'HubSpot Contact'."},"type":{"type":"string","description":"Custom field type: TEXT, NUMBER, PRICE, DATE, DATETIME, or EXTERNAL_ID. Defaults to TEXT."},"config":{"type":["object","string","null"],"description":"Type-specific config, as an object (preferred) or a JSON string. EXTERNAL_ID requires {url_template:'https://.../{value}'}; PRICE requires {currency:'USD'}; DATE/DATETIME may set {format}."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"if_not_exists":{"type":"boolean","description":"Default true. If a custom field with the same name already exists, return it instead of creating a duplicate."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["id","name","type","mapping_value","existed"],"properties":{"id":{"type":"string"},"name":{"type":"string"},"type":{"type":"string"},"config":{"type":["object","null"]},"existed":{"type":"boolean","description":"True when if_not_exists reused an existing custom field."},"mapping_value":{"type":"string","description":"Wire mapping value to use in import mappings, e.g. CUSTOM.123."}}}},{"name":"leadbay_update_custom_field","annotations":{"title":"Update CRM custom field","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":true},"description":"$a6","inputSchema":{"type":"object","required":["id","_triggered_by"],"properties":{"id":{"type":"string","description":"Custom field id to update (the numeric id from leadbay_list_mappable_fields, NOT the 'CUSTOM.<id>' mapping value)."},"name":{"type":"string","description":"New user-visible name. Omit to keep the current name."},"type":{"type":"string","description":"New type: TEXT, NUMBER, PRICE, DATE, DATETIME, or EXTERNAL_ID. Omit to keep the current type."},"config":{"type":["object","string","null"],"description":"New type-specific config, as an object (preferred) or a JSON string. EXTERNAL_ID requires {url_template:'https://.../{value}'}; PRICE requires {currency:'USD'}; DATE/DATETIME may set {format}. Omit to keep current config."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["id","name","type","mapping_value"],"properties":{"id":{"type":"string"},"name":{"type":"string"},"type":{"type":"string"},"config":{"type":["object","null"]},"mapping_value":{"type":"string","description":"Wire mapping value for import mappings, e.g. CUSTOM.123."}}}},{"name":"leadbay_delete_custom_field","annotations":{"title":"Delete CRM custom field","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":true},"description":"$a7","inputSchema":{"type":"object","required":["id","_triggered_by"],"properties":{"id":{"type":"string","description":"Custom field id to delete (the numeric id from leadbay_list_mappable_fields, NOT the 'CUSTOM.<id>' mapping value)."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"confirm":{"type":"boolean","description":"Must be true to actually delete. Without it the tool returns the field that WOULD be deleted and does nothing — a safety gate, because deletion drops the field's values from every lead."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["id","deleted"],"properties":{"id":{"type":"string"},"hint":{"type":"string","description":"Guidance — present when confirm was missing (re-call with confirm:true)."},"name":{"type":"string"},"type":{"type":"string"},"deleted":{"type":"boolean","description":"True when the field was deleted; false when confirm was not set."}}}},{"name":"leadbay_set_qualification_questions","annotations":{"title":"Modify the org's qualification questions","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":true},"description":"$a8","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"add":{"type":"array","items":{"type":"string","maxLength":255},"description":"Questions to append to the current list (deduped). Mutually exclusive with `questions`. A question Leadbay can score has ALL of: the estimative marker — it starts \"Is the company likely to \" in English or \"L'entreprise est-elle susceptible de/d' \" in French (never \"Does the company ...\" or a bare \"L'entreprise est-elle <X> ?\", which the scorer cannot verify and marks no); ONE dimension, not two joined by AND; something estimable from the company's public material, never its budget or its internal plans; and enough bite to split companies roughly 30/70 — a question nearly everyone answers yes to (\"has a website\") adds no signal, so say that and propose a sharper one rather than writing it. Max 120 chars, in the user's language. Never name a specific company in a question. Read leadbay_get_qualification_questions first: skip anything an existing question already tests, and anything the lens already filters by sector, headcount or territory."},"remove":{"type":"array","items":{"type":"string"},"description":"Exact question strings to remove from the current list. Mutually exclusive with `questions`. A removal requires confirm:true."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"confirm":{"type":"boolean","description":"Required whenever the change DROPS ANY existing question — including a same-count swap or a `questions` replacement that omits a current question, not only when the list gets shorter (removing a question changes how every lead is scored). Without it, such a change is previewed and not applied. Pure additions never need confirm."},"questions":{"type":"array","items":{"type":"string","maxLength":255},"description":"$a9"},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"add_anti_patterns":{"type":"array","items":{"type":"string"},"description":"Negative criteria to append to the org's ideal buyer profile: kinds of company that are never the buyer (\"Consulting firms and IT services companies\", \"Franchise locations of national chains\"). Qualification reads them as negative signals when it scores a lead. Uses no question slot. Its own write: do not combine with questions/add/remove. Admin only. Read leadbay_get_qualification_questions first and skip anything an existing anti-pattern or question already says. In the user's language."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["qualification_questions","count","changed"],"properties":{"hint":{"type":"string","description":"Operator note — confirm prompt on a shrink, or a no-op explanation."},"_meta":{"type":"object"},"count":{"type":"number"},"region":{"type":"string"},"changed":{"type":"boolean","description":"True when the list was actually written; false on a no-op or an unconfirmed shrink."},"anti_patterns":{"type":"array","items":{"type":"string"},"description":"The ideal buyer profile's negative criteria after an add_anti_patterns call."},"form_warnings":{"type":"array","items":{"type":"string"},"description":"Present when a written question will score poorly — not in the estimative 'Is the company likely to ...' form, or over 120 chars. The change WAS applied; tell the user and offer the rewrite."},"previous_count":{"type":"number"},"anti_patterns_added":{"type":"array","items":{"type":"string"},"description":"The criteria this call wrote."},"qualification_questions":{"type":"array","items":{"type":"object"},"description":"The questions AFTER the change. Each: {question}."}}}},{"name":"leadbay_add_note","annotations":{"title":"Add a note on a lead","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":false},"description":"Add a note to a lead. Notes are visible to the whole organization in Leadbay.\n\nWHEN TO USE: low-level — for free-form notes not tied to outreach actions, including meaningful per-lead notes/context preserved from an imported file after the import returns lead IDs. Also to record why a company was set aside when a CRM status change carries no reason, or when the user tells you something about one company that only they know (their own sales agent, already a client) without excluding it. A dislike saves its own reason: do not add a second note for it.\n\nWHEN NOT TO USE: to log an outreach action — use leadbay_report_outreach, which requires verification (gmail/calendar/user_confirmed) to prevent hallucinated outreach poisoning the SDR pipeline.\n\nThis tool MUTATES state. The caller (agent or human-in-the-loop) is responsible for confirming intent before invocation; the MCP server does not soft-prompt for confirmation. See `annotations.destructiveHint`.\n","inputSchema":{"type":"object","required":["leadId","note"],"properties":{"note":{"type":"string","description":"Note text (max 4095 characters)"},"leadId":{"type":"string","description":"Lead UUID (required)"},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["id","note","created_at"],"properties":{"id":{"type":"string","description":"Note id assigned by the backend."},"note":{"type":"string","description":"Echoed note text (truncated to 4095 chars)."},"created_at":{"type":"string","description":"ISO timestamp of creation."}}}},{"name":"leadbay_like_lead","annotations":{"title":"Like a lead","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$aa","inputSchema":{"type":"object","required":["lead_id"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"lead_id":{"type":"string","description":"UUID of the lead to like."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}},{"name":"leadbay_dislike_lead","annotations":{"title":"Dislike a lead","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$ab","inputSchema":{"type":"object","required":["lead_id"],"properties":{"reason":{"type":"string","description":"Why the lead is rejected, in the user's own words (\"société de conseil, hors ICP\", \"unpaid invoice with them\"). In a run with no user, the criterion you applied. Saved as a note on the lead that the whole team sees. Omit when no reason was given; never invent one."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"lead_id":{"type":"string","description":"UUID of the lead to dislike."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}},{"name":"leadbay_set_lead_status","annotations":{"title":"Set lead CRM status","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$ac","inputSchema":{"type":"object","required":["lead_ids","status"],"properties":{"status":{"type":"string","description":"One of: WANTED, WON, LOST, UNWANTED (case-insensitive). DEFAULT and INBOUND are accepted but are normally set by Leadbay itself."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"lead_ids":{"type":"array","items":{"type":"string"},"description":"Lead UUIDs (1-200). Every lead gets the same status."},"status_date":{"type":"string","description":"Optional YYYY-MM-DD — the date the status was actually reached (close date). Omit to let the backend stamp now."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}},{"name":"leadbay_create_campaign","annotations":{"title":"Create a named campaign (optionally seeded with leads)","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":false},"description":"$ad","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"name":{"type":"string","description":"Campaign display name (max 255 chars). Omit to let the backend AI-generate one from the seed lead_ids. If lead_ids is empty AND name is omitted, the backend assigns a default."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"lead_ids":{"type":"array","items":{"type":"string"},"description":"Lead UUIDs to attach at creation. Empty array (default) creates an empty campaign — add leads later via leadbay_add_leads_to_campaign. Non-empty seed enables AI name suggestion."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["id","name","created_at"],"properties":{"id":{"type":"string","description":"Campaign UUID."},"name":{"type":"string"},"archived":{"type":"boolean"},"created_at":{"type":"string"},"created_by":{"type":"string"},"updated_at":{"type":"string"},"ai_name_count":{"type":"number"},"last_accessed_at":{"type":"string"},"ai_generated_name":{"type":["string","null"]}}}},{"name":"leadbay_add_leads_to_campaign","annotations":{"title":"Add leads to an existing campaign","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$ae","inputSchema":{"type":"object","required":["campaign_id","lead_ids","_triggered_by"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"lead_ids":{"type":"array","items":{"type":"string"},"minItems":1,"description":"Lead UUIDs to add. Backend rejects unknown lead UUIDs with 404 — pass UUIDs sourced from pull_leads / pull_followups / tour_plan / research."},"campaign_id":{"type":"string","description":"Campaign UUID (from leadbay_create_campaign or leadbay_list_campaigns)."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["added","already_present"],"properties":{"added":{"type":"number","description":"Leads newly attached."},"already_present":{"type":"number","description":"Leads that were already in this campaign — no-op."}}}},{"name":"leadbay_remove_leads_from_campaign","annotations":{"title":"Remove leads from a campaign","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":true},"description":"$af","inputSchema":{"type":"object","required":["campaign_id","lead_ids","_triggered_by"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"lead_ids":{"type":"array","items":{"type":"string"},"minItems":1,"description":"Lead UUIDs to remove. Pass IDs sourced from Leadbay tools; invalid IDs are handled by the backend."},"campaign_id":{"type":"string","description":"Campaign UUID (from leadbay_create_campaign or leadbay_list_campaigns)."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["removed"],"properties":{"removed":{"type":"number","description":"Number of leads submitted for removal (backend returns 204, no per-lead breakdown)."}}}},{"name":"leadbay_extend_lens","annotations":{"title":"Extend a lens with additional leads (extra refill)","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":false},"description":"$b0","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"lensId":{"type":"number","description":"Lens to extend. Defaults to the user's last-active lens."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"extra_count":{"type":"number","description":"How many extra leads to request. Omit to use the backend default. The full requested count must fit into the remaining daily LENS_EXTRA_REFILL quota — otherwise the call is rejected outright (status: quota_exceeded)."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"seed_lead_ids":{"type":"array","items":{"type":"string"},"description":"Optional list of lead UUIDs from leadbay_list_lens_seed_candidates to bias the recommender. Omit or empty array → default-strategy fallback (same behaviour as a normal fill)."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["status","lens"],"properties":{"lens":{"type":"object","properties":{"id":{"type":"number"}}},"quota":{"type":"object","properties":{"resets_at":{"type":["string","null"]},"used_today":{"type":["number","null"]}},"description":"Only present on status=quota_exceeded. Shows the org's daily LENS_EXTRA_REFILL state."},"reason":{"type":"object","required":["code","message","retryable"],"properties":{"code":{"type":"string","description":"audience_too_narrow (lens holds nothing and its criteria intersect to nothing) | no_new_leads (everything matching has already been delivered) | no_candidates (pool is empty; which of the two could not be determined)"},"message":{"type":"string"},"criteria":{"type":"object","description":"The criteria in play, so the agent can name them."},"retryable":{"type":"boolean","description":"Always false here. Re-calling leadbay_extend_lens cannot change the outcome — widen the audience instead."},"narrow_locations":{"type":"array","items":{"type":"object"},"description":"Include-locations that resolved to city-scale or smaller — the usual thing to relax first."}},"description":"Only present on status=no_candidates. Same shape and `code` vocabulary as leadbay_pull_leads' empty_reason, so one routing rule covers both tools."},"status":{"type":"string","description":"queued | no_candidates | quota_exceeded | refresh_in_progress | no_valid_seeds"},"message":{"type":"string","description":"Human-readable summary. On error statuses, this is the line to surface to the user."},"accepted_seeds":{"type":"array","items":{"type":"string"},"description":"Subset of seed_lead_ids that passed validation and will bias the fill. Empty when no seeds were submitted (default-strategies fallback). Present only on status=queued."},"available_count":{"type":["number","null"],"description":"How many leads a refill on this lens could still draw, read from /extra_refill_preview before queueing. 0 means the refill was NOT queued (status=no_candidates). null means the pool could not be read and the refill was queued anyway."}}}},{"name":"leadbay_manage_lenses","annotations":{"title":"List, switch, edit, or delete your lenses","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":true},"description":"$b1","inputSchema":{"type":"object","required":["_triggered_by"],"properties":{"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"confirm":{"type":"boolean","description":"Required (=true) to actually delete. Without it, deleteLensId returns a preview to confirm with the user first."},"newName":{"type":"string","description":"New lens name (used with editLensId)."},"editLensId":{"type":["string","number"],"description":"When set, edit this lens's metadata — provide newName and/or newDescription. Must be one of the user's lenses."},"deleteLensId":{"type":["string","number"],"description":"When set, delete this lens. DESTRUCTIVE — returns a delete_preview unless confirm:true. Cannot delete the default lens."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"newDescription":{"type":"string","description":"New lens description (used with editLensId). Pass an empty string to clear it."},"switchToLensId":{"type":["string","number"],"description":"When set, switch the active lens to this id (must be one of the user's lenses), then return the refreshed list."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["status","lenses","active_lens_id"],"properties":{"edited":{"type":"boolean","description":"True when this call renamed/re-described a lens."},"lenses":{"type":"array","items":{"type":"object"},"description":"The user's lenses, each with every field GET /lenses returns plus is_active, is_default and `criteria` (its sectors, locations and sizes by name; null if unreadable)."},"status":{"type":"string","description":"'listed', 'switched', 'already_active', 'edited', 'deleted', 'delete_preview' (confirm to proceed), 'cannot_delete_default', or 'not_found'."},"deleted":{"type":"boolean","description":"True when this call deleted a lens."},"message":{"type":"string"},"switched":{"type":"boolean","description":"True when this call changed the active lens."},"will_delete":{"type":"object","description":"On 'delete_preview': the lens that WILL be deleted {id, name}. Nothing removed yet."},"active_lens_id":{"type":["string","null"]}}}},{"name":"leadbay_new_lens","annotations":{"title":"Create a new named lens","readOnlyHint":false,"openWorldHint":false,"idempotentHint":false,"destructiveHint":false},"description":"$b2","inputSchema":{"type":"object","required":["name","_triggered_by"],"properties":{"base":{"type":"number","description":"Lens id to clone from. Defaults to the active/default lens."},"name":{"type":"string","maxLength":255,"description":"Display name for the new lens (required)."},"sizes":{"type":"array","items":{"type":"object","properties":{"max":{"type":"number"},"min":{"type":"number"}}},"description":"Company size buckets, e.g. [{min:30,max:300}]."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"confirm":{"type":"boolean","description":"Safety gate. Defaults to false → the tool returns a PREVIEW and creates nothing. Show the preview to the user, get their explicit go-ahead, then re-call the SAME args with confirm:true to actually create the lens."},"sectors":{"type":"array","items":{"type":"string"},"description":"Sectors to include — free text (auto-resolved) or ids."},"locations":{"type":"array","items":{"type":"string"},"description":"Geographic scope — free text (e.g. ['Indre-et-Loire', 'Texas']) or admin-area ids. Resolved via /geo/search at any level from state down to city (state / région / département / county / city). NEVER a country name — this workspace serves exactly ONE country, so a whole-country ask means passing NO location at all (rejected with COUNTRY_LEVEL_LOCATION). Scopes the lens to a sales territory."},"description":{"type":"string","maxLength":255,"description":"Optional lens description."},"_triggered_by":{"type":"string","description":"MANDATORY — copy/paste the verbatim portion of the user's most recent message that this call is acting upon. Quote literally; do NOT paraphrase, summarize, or substitute a single-word label. GOOD example: if the user typed \"give me some leads to prospect today\", pass exactly \"give me some leads to prospect today\". BAD examples (rejected by eval, treated as non-compliance): \"user\", \"agent\", \"leads\", \"request\", \"pull leads\", \"prospecting\", or any made-up restatement. If you are acting WITHOUT a fresh user message (a memory recall, a scheduled run, a self-initiated retry), pass the actual instruction you are acting on — the recalled directive, the schedule's intent, or the original request being retried — so the value is always a real, auditable trace. Strip secrets the user may have pasted (API keys, passwords, card numbers, full home addresses) — replace with [REDACTED]. The call is rejected as LAST_PROMPT_REQUIRED if missing or blank."},"exclude_sectors":{"type":"array","items":{"type":"string"},"description":"Sectors to exclude — free text or ids."},"exclude_locations":{"type":"array","items":{"type":"string"},"description":"Locations to exclude — free text or ids. Sub-country areas only — excluding a country is meaningless on a single-country workspace and is rejected."}},"additionalProperties":false},"outputSchema":{"type":"object","required":["status"],"properties":{"lens":{"type":"object","description":"On 'created': the created lens {id, name}."},"_meta":{"type":"object"},"status":{"type":"string","description":"'preview', 'created', 'ambiguous_sectors', 'ambiguous_locations', 'country_level_location', or 'orphan_created' (filter write failed + cleanup failed)."},"message":{"type":"string"},"will_create":{"type":"object","description":"On 'preview': what WILL be created — {name, description, sectors, exclude_sectors, sizes, locations, exclude_locations}. Nothing has been written yet."},"filter_applied":{"type":"object","description":"On 'created': the FilterPayload POSTed to the new lens."},"filter_preview":{"type":"object","description":"On 'preview': the FilterPayload that would be applied."},"country_locations":{"type":"array","items":{"type":"object"},"description":"On 'country_level_location': per offending value {value, param, kind, country, axis, kept}. A country name is never a location criterion — each workspace serves exactly ONE country. The recovery BRANCHES on `country_locations[].axis` and `[].kind`; `hint` states the one for THIS call — follow it verbatim. When the country was the ONLY scope, or on ANY non-foreign `exclude`, the answer is to write NOTHING at all — re-calling with the value merely dropped persists a scope that inverts the request. Never retry with another spelling or a nearby city."},"computing_wishlist":{"type":"boolean","description":"On 'created': true when the lens has criteria and its wishlist is (re)computing asynchronously — an immediate leadbay_pull_leads may read empty; wait ~30s. False for a criteria-less clone (it inherits the base lens's leads immediately)."},"sector_ambiguities":{"type":"array","items":{"type":"object"},"description":"On 'ambiguous_sectors': per text {sector_text, matches:[{id,name,score}]}."},"location_ambiguities":{"type":"array","items":{"type":"object"},"description":"On 'ambiguous_locations': per text {location_text, matches:[{id,name,country,level,score}]}. Re-call the chosen id via the SAME axis the text came from — an include text → locations; a text from exclude_locations → exclude_locations (NOT locations, which would include the area the user asked to exclude). The `message` field names the correct param per text."}},"description":"'preview' (default, NOTHING created — confirm with the user then re-call with confirm:true); 'created' on success; 'ambiguous_sectors' / 'ambiguous_locations' when free-text sectors / locations didn't resolve (re-call with ids — the lens was NOT created); 'country_level_location' when a country-level value was passed as a location (the lens was NOT created; read `hint` — re-calling without the value is often itself wrong)."}},{"name":"leadbay_set_telemetry","annotations":{"title":"Enable, disable, or check product-usage telemetry","readOnlyHint":false,"openWorldHint":false,"idempotentHint":true,"destructiveHint":false},"description":"$b3","inputSchema":{"type":"object","properties":{"action":{"enum":["enable","disable","status"],"type":"string","description":"enable / disable flip telemetry for the user; status just reports the current setting. Defaults to status."},"_origin":{"enum":["agent","artifact"],"type":"string","description":"OPTIONAL METADATA — set to \"artifact\" ONLY by the Leadbay artifact runtime when a control inside a rendered artifact makes this call. Agents must NOT set this field; an agent-issued call is recorded as \"agent\" by omitting it. Does not affect tool behavior."},"_triggered_by":{"type":"string","description":"OPTIONAL METADATA — the verbatim user utterance (or short paraphrase) that led you to call this tool. Pass the user's literal phrasing (last 1-3 sentences). Records what the call is acting upon for context and audit. Does not affect tool behavior. Always include when you have it."}},"additionalProperties":false}}],"prompts":[{"name":"leadbay_daily_check_in","arguments":[],"description":"Morning DISCOVERY workflow — new leads from the lens wishlist. Trigger\non \"show me leads\", \"what's new today\", \"let's prospect\", \"run my check-in\",\n\"my morning check-in\", \"I do this every day\", \"every morning\". Recurrence\nlanguage always means this prompt. Do NOT trigger on follow-up phrasings\n(\"follow up\", \"before my trip\") — those go to `leadbay_followup_check_in`.\n"},{"name":"leadbay_prospecting_overview","arguments":[],"description":"Orientation for working with Leadbay from any host — discovery vs.\nfollow-up, the outreach loop, outcome recording, imports, pushback /\nsnooze, and the connected-outreach-tool registry. Trigger when the\nconversation involves Leadbay leads, prospecting, pipeline, follow-up,\noutreach, or lens / ICP — anything from \"show me my leads\" to \"what\nshould I follow up on\" to \"I'll send via lemlist\".\n"},{"name":"leadbay_new_leads","arguments":[{"name":"need","required":false,"description":"What the user is looking for, in their own words (e.g. '10 gyms around Dallas that would buy modular flooring, with phone numbers'). Optional — the session starts by asking when absent."}],"description":"Guided net-new lead delivery — turn a described need (\"gyms around Dallas\nthat would buy our flooring\") into ICP-perfect NEW companies with\nqualification evidence and the right contact, via leadbay_find_new_leads.\nTrigger when the user DESCRIBES who they want: \"get me N companies that\n<profile>\", \"we're entering <market>\". A bare \"find me new leads\" with no\nprofile, and \"today's leads\", are the daily lens batch — leadbay_pull_leads.\n\"Qualify these companies I have\" is leadbay_qualify_leads.\n"},{"name":"leadbay_research_a_domain","arguments":[{"name":"domain","required":true,"description":"Company name or domain (for example 'Acme Corporation' or 'acme.com'). The legacy argument key remains `domain` for client compatibility."}],"description":"Resolve a company by name or domain across the user's visible Discover,\nMonitor, and Activate corpus, then return everything Leadbay knows about it.\n"},{"name":"leadbay_import_file","arguments":[{"name":"file","required":false,"description":"Path or user-visible name of the CSV/file to import. If omitted, use the file the user attached or referenced."},{"name":"instruction","required":false,"description":"Additional user goal, e.g. \"then qualify the leads\", \"preserve owner phone as a custom field\", or \"only import restaurants in Manhattan\"."}],"description":"Import a user-supplied CSV/file into Leadbay through five phases with\nevidence gates — scan, derive, resolve identities, preserve & commit,\nthen optionally qualify and report. The job is to maximize how many\nrows the Leadbay system actually ingests and matches.\n"},{"name":"leadbay_refine_audience","arguments":[{"name":"instruction","required":true,"description":"The refinement (e.g. 'focus on hospitals running their own IT'). Set to plain English."}],"description":"Refine the kind of leads Leadbay surfaces beyond firmographics, with a\nfree-text instruction. Handles the clarification round-trip if the new\nprompt is ambiguous.\n"},{"name":"leadbay_log_outreach","arguments":[{"name":"lead_id","required":true,"description":"The lead UUID. Get it from leadbay_pull_leads or leadbay_research_lead_by_id."},{"name":"summary","required":true,"description":"1-2 sentences describing what I did (e.g. 'Sent intro email to CTO citing recent Hornsea contract')."}],"description":"Log outreach (an email I sent, a call I made, a meeting I had) on a\nspecific lead. Captures verification so the SDR pipeline trusts the entry.\n"},{"name":"leadbay_sync_outreach","arguments":[{"name":"lookback_days","required":false,"description":"Optional: how many days back the first run reads (default 14). The daily task always reads the last 2 days."}],"description":"Keep Leadbay's record of who was contacted current: read my mailbox and\ncalendar, log every email and meeting with a Leadbay lead on the person it\ninvolved, then schedule the same sync to run every day. Trigger on \"log my\nemails in Leadbay\", \"sync my outreach\", \"keep Leadbay up to date with my\nemails\".\n"},{"name":"leadbay_plan_tour_in_city","arguments":[{"name":"city","required":true,"description":"City or region the user is visiting (e.g. 'Limoges', 'Bay Area'). Used as the geo filter for both Monitor and Discover lookups. A country is not a city: this workspace already covers exactly one country, and a country name here silently fences the tour to a same-named village. Do NOT omit the argument to recover — a city-less tour returns arbitrary leads from across the whole workspace, which is not an itinerary. Ask which city or region the visit is to."},{"name":"date","required":false,"description":"When the visit is (e.g. 'May 24', 'next Thursday'). Surfaced in the outreach drafts as 'I'll be in <city> on <date>'."}],"description":"Use whenever the user names a city they'll be in and asks who to see\n— \"I'm in SF next Tuesday, who's worth meeting?\", \"I'm going to Berlin\n— who should I visit?\", \"plan my <city> tour\". Any in-person/visit\nintent tied to a place routes here, NOT to `leadbay_pull_leads`. It\nsurfaces follow-ups + fresh Discover leads in the city via\n`leadbay_tour_plan`, ALWAYS offers to plot them on a map (rendering it\non yes), then offers outreach drafts + campaign persistence.\n"},{"name":"leadbay_build_campaign","arguments":[{"name":"audience","required":false,"description":"Optional: a fresh audience to target (e.g. 'dental clinics in Texas'). Omit to build from your ACTIVE lens — the default."},{"name":"campaign_name","required":false,"description":"Optional: a name for the campaign. Omit and one is derived from the lens/audience + date (or the backend AI-names it)."},{"name":"count","required":false,"description":"Optional: how many fully-actionable leads to build (default 20). The loop keeps discovering, qualifying and enriching until this many in-ICP leads each have a reachable target-title contact — or the lens is exhausted. Higher counts take longer and consume more quota."},{"name":"job_titles","required":false,"description":"Optional: the exact buyer job titles to enrich, comma-separated (e.g. 'VP Sales, Head of Growth, Director of Business Development'). Omit and the buyer persona is derived from what you sell. A lead only counts toward the target when it has a reachable contact matching one of these titles."}],"description":"Build a sales campaign from scratch, autonomously, to a target size:\ndiscover on the lens, qualify, and enrich the buyer titles until `count`\nleads each have a reachable target-title contact — no pauses, no confirm\ngates. Saves via `leadbay_create_campaign` and renders a one-tap\ncall/email view via `leadbay_campaign_call_sheet`. Trigger on \"build me a\ncampaign\", \"build N leads\", \"create a campaign from scratch\". Work an\nexisting one with `leadbay_work_campaign`.\n"},{"name":"leadbay_setup_team_prospecting","arguments":[{"name":"audience","required":true,"description":"Natural-language audience description (e.g. 'plumbing companies with 10-50 employees in Seine-Maritime'). The lens-creation step (`leadbay_refine_lead_targeting` → `leadbay_create_lens`) interprets it. A country name is not a scope here — this workspace already covers exactly one country, so drop it and keep the rest of the description; a DIFFERENT country cannot be targeted at all."},{"name":"rep_split","required":false,"description":"Optional: how to split the validated leads into per-rep campaigns. Free text — e.g. 'split by city' or 'one campaign per rep: John gets Tulsa, Sarah gets OKC'. Splitting by country is not a split — the workspace is single-country."}],"description":"Manager-led prospecting setup: conversationally turn a natural-language\naudience ask into a Leadbay lens, validate the candidate leads, and\npersist them as one or more named campaigns the rep(s) can work\nthrough. Closes #3630 US3 end-to-end (within the current\ncreator-scoped campaign visibility model).\n"},{"name":"leadbay_work_campaign","arguments":[{"name":"campaign","required":false,"description":"Campaign name (fuzzy match against your own campaigns) or campaign UUID. Omit to list and pick interactively."},{"name":"mode","required":false,"description":"Optional: skip the readiness-assessment proposal and jump directly into 'call_sheet' / 'email_sheet' / 'map' / 'enrich_first'. Omit (recommended) and let the prompt propose based on the data."}],"description":"Work a campaign as a real outreach session: pick the campaign,\nassess what the user has (phones / emails / coords), then PROPOSE\nthe right session mode (call sheet, email sheet, enrich titles\nfirst, map). After they pick, render — and as they dictate\noutcomes per lead, record both note + epilogue via\n`leadbay_report_outreach` in one round trip.\n"},{"name":"leadbay_outreach_session","arguments":[{"name":"source","required":false,"description":"Optional: where the leads come from — 'followups' (the Monitor, default), 'discover' (the active lens), or 'campaign'. Omit and the session asks."},{"name":"campaign","required":false,"description":"Optional: campaign name (fuzzy-matched against your own campaigns) or UUID, when source is 'campaign'. Omit to list and pick."}],"description":"Run an outreach session over ANY set of leads: follow-ups, a Discover\nlens, or a campaign. Assess who is actually reachable, build the lead\ndesk, then record each outcome as you dictate it. Trigger on \"relance\nplan\", \"campagne de relance\", \"cold calling session\", \"prospection\nsession\", \"outreach session\", \"let's contact leads\", \"who do I call\ntoday\". For a campaign the user NAMES, prefer `leadbay_work_campaign`.\n"},{"name":"leadbay_route_session","arguments":[{"name":"city","required":false,"description":"Where the rep is going — a city or region (e.g. 'Lyon', 'Hauts-de-France'). Omit and the session asks. NEVER a country: this workspace serves one country, so a country name is not a place to tour."},{"name":"day","required":false,"description":"Optional: which day the trip is, in the rep's own words ('Thursday', 'next week'). Used only to frame the plan; nothing is scheduled."}],"description":"Plan and work a day on the road: pick the place, see the leads on a map,\ndrive them in a sensible order and record each visit as it happens.\nTrigger on \"plan my route\", \"tournée\", \"road trip\", \"I'm driving to\n<city>\", \"who can I visit in <city>\", \"plan my day in the field\",\n\"prospection terrain\". For working leads by PHONE from a desk, prefer\n`leadbay_outreach_session`.\n"},{"name":"leadbay_qualify_top_n","arguments":[{"name":"count","required":false,"description":"How many leads to qualify (default 10, max 25). Higher counts may take 5+ minutes."}],"description":"Bulk-qualify the top N un-qualified leads in the active lens. Uses\nleadbay_bulk_qualify_leads with a sensible default budget.\n"},{"name":"leadbay_top_accounts_to_activate","arguments":[{"name":"count","required":false,"description":"Optional: how many accounts the plan should hold (default 50)."},{"name":"territory","required":false,"description":"Optional: restrict the plan to a territory (e.g. 'Indre-et-Loire', 'Région Ouest'). Sets geography on the Discover lens. A country is not a territory — this workspace already covers exactly one country."}],"description":"Build a ranked account-conquest plan from Leadbay data — the accounts worth\nactivating, each with a motif, a pitch and a checklist, ranked by the\nstrongest Leadbay signal. Every figure carries its source, and anything\nLeadbay can't measure is shown as OMITTED rather than estimated. Uses\n`leadbay_bulk_qualify_leads` and `leadbay_enrich_titles`. Trigger on\n\"top 50 accounts to activate\", \"who should we go after\".\n"},{"name":"leadbay_getting_started","arguments":[],"description":"Guided first-run walkthrough — four clicks that actually use Leadbay: check\nthe account, pull today's leads, draft a first email to the top one, then\nreveal who to send it to. Use when the user is new or asks to be SHOWN how\nLeadbay works (\"walk me through Leadbay\", \"I'm new\", \"how do I use this\",\n\"give me a tour\"). Don't use it for orientation prose with no clicking —\nthat's leadbay_prospecting_overview.\n"}],"resources":[{"uri":"org://taste-profile","name":"Org taste profile","mimeType":"application/json","description":"The org's qualification questions, intent tags, and ICP signals — the agent's knowledge base for what makes a lead a fit."}],"triggers":[],"configSchema":null,"configAllowsEmpty":false},"className":"w-12 h-12 shrink-0"}],"$Lb4"]}],"$Lb5"]}],"$Lb6"]}]}],"$Lb7"]}],"$Lb8"]}]] </article></body></html>