reflect-yourself

A self-learning system for Cursor that captures corrections and syncs them to the right place.

Workflow

When invoked:

1. Analyze session (and optionally ask clarifying questions if scope is unclear)
2. Present learnings for review (summary-first, then cards; see command file)
3. Capture action via the Cursor agent tool Ask questions (message-question) when possible — offer Apply all / Apply selected / Skip all / Partial / Other; see Cursor docs → Agent overview → Tools
4. Apply only after explicit user approval (via Q&A or typed reply)

Inputs

Commands

When the user invokes these commands, read the corresponding file in commands/:

Quick Start

1. At session end, run /reflect-yourself
2. Review captured learnings (summary + cards)
3. Choose action via prompt or reply (apply / edit / skip / redirect)
4. Learnings are applied to skills/rules after approval

Placement Logic

• Project Skills (.cursor/skills/) → Codebase-specific
• Personal Skills (~/.cursor/skills/) → Reusable across projects
• Project Rules (.cursor/rules/) → Coding standards

Queue Location

Pending learnings: ~/.cursor/reflect-queue.json (global location, never pollutes project repos)

Security

• Trust model: Only install skills from sources you trust. Skills are loaded into the agent context and can influence behavior.
• Import hygiene: Prefer official or well-known repos; inspect SKILL.md and command/rules files before installing. Be wary of skills that instruct the agent to run arbitrary code or exfiltrate data.
• This skill: Only instructs the agent to write to .cursor/skills/, .cursor/rules/, and ~/.cursor/skills/. The installer copies files only; it does not execute skill content. After install, review ~/.cursor/skills/reflect-yourself/ if you want to verify contents.

Update check

When main commands are run, the agent may check for a newer npm version. If one exists, the user can be prompted (Yes/No) to run npx reflect-yourself@latest. See command files for details.