Smithery Logo
MCPsSkillsDocsPricing
Login
NewFlame, an assistant that learns and improves. Available onTelegramSlack
    xingmiao201081

    traecnclaw-mcp

    xingmiao201081/traecnclaw-mcp

    About

    SKILL.md

    Install

    • Telegram
      Telegram
    • Slack
      Slack
    • Claude Code
      Claude Code
    • Codex
      Codex
    • OpenClaw
      OpenClaw
    • Cursor
      Cursor
    • Amp
      Amp
    • GitHub Copilot
      GitHub Copilot
    • Gemini CLI
      Gemini CLI
    • Kilo Code
      Kilo Code
    • Junie
      Junie
    • Replit
      Replit
    • Windsurf
      Windsurf
    • Cline
      Cline
    • Continue
      Continue
    • OpenCode
      OpenCode
    • OpenHands
      OpenHands
    • Roo Code
      Roo Code
    • Augment
      Augment
    • Goose
      Goose
    • Trae
      Trae
    • Zencoder
      Zencoder
    • Antigravity
      Antigravity
    • Download skill
    ├─
    ├─
    └─
    Smithery Logo

    Give agents more agency

    Resources

    DocumentationPrivacy PolicySystem Status

    Company

    PricingAboutBlog

    Connect

    © 2026 Smithery. All rights reserved.

    About

    Use when an MCP-capable agent must connect, configure, operate, test, or document TRAECNclaw's stdio MCP server for TraeCN automation, OpenClaw integration, unattended vibe workflows,...

    SKILL.md

    TRAECNclaw MCP

    Drive TraeCN through its real stdio MCP server, backed by the local HTTP gateway and CDP automation. Client-neutral: works for any agent that can call stdio MCP tools.

    When to use

    • Need to run, queue, poll, or cancel TraeCN tasks
    • Switch model or Solo/IDE mode, read/write TraeCN settings
    • Handle confirm dialogs or review gates
    • Run unattended vibe workflows or long-queue proofs
    • Configure any MCP client (Cursor, Claude Desktop, Cline, Windsurf, Codex)

    Do NOT use this skill for direct browser automation or GUI clicking when an MCP tool exists.

    Workflow

    1. Confirm Node 22+ and the gateway can reach TraeCN.
    2. Discover before acting: tools/list → traecn_get_capabilities → traecn_preflight with the intended command/params.
    3. For side effects, honor the traecn_preflight or traecn_plan_operation result (confirmation, readiness, queue, dialog) before operating.
    4. Use traecn_list_commands + traecn_run_command only when a client needs the stable command catalog rather than per-tool MCP names.
    5. Start the gateway separately: npm run start:gateway (or TRAECN_ENABLE_MOCK_BRIDGE=1 npm run start:gateway for mock mode).
    6. For exact tool names, profiles, and MCP client config JSON, read references/mcp-surface.md.

    Start the MCP server

    Use the bundled launcher so the server inherits stdio for JSON-RPC:

    node scripts/start-mcp.js
    

    The launcher auto-locates mcp-server.js from the repo root. Set TRAECN_MCP_SERVER_PATH to override. For MCP client config templates (Cursor, Claude Desktop, Cline, Windsurf), copy assets/mcp-client-config.json and replace SKILL_DIR with the absolute path to this skill folder.

    Profile choice

    TRAECN_MCP_TOOL_PROFILE scopes tools/list:

    Profile Tools Use
    public 20 Default for published agent integrations
    ops 32 Recovery, cleanup, Solo control, long-queue proof
    full 36 Audits, compatibility checks, every explicit shortcut

    The full command catalog stays reachable via traecn_list_commands / traecn_run_command regardless of profile.

    Safety rules

    • Keep the gateway bound to 127.0.0.1 unless remote access is intentionally required.
    • Never commit .env, tokens, auth headers, raw TraeCN profile data, audit-logs/, demo-output/, logs/, or node_modules/.
    • Prefer MCP/HTTP/CLI surfaces over GUI-control paths whenever the MCP server can perform the operation.
    • Treat live dialog text, gateway health, cleanup state, and external model queue behavior as environment-specific until revalidated.

    Validation

    • npm run doctor — local environment readiness.
    • npm test — maintained in-repo test sequence.
    • npm run acceptance:audit -- --require-complete — before claiming unattended end-to-end readiness.
    • npm run test:all — release-style changes when practical.
    • node -e "console.log(require('./mcp-server').MCP_TOOLS.map(t => t.name).join('\n'))" — verify live tool list from code.
    Recommended Servers
    MCP Hive
    MCP Hive
    agent-advisories
    agent-advisories
    Repository
    luckycat133/traecnclaw-mcp-skill
    Files