
Self-Improving Agent
Curates Claude Code's auto-memory into permanent project rules and reusable skills
- Repository
- alirezarezvani/claude-skills
- Stars
- 29,149
- Licence
- MIT
- Files
- 1 file
What it does
It helps an AI coding agent review what it has automatically noted about your project (stored in MEMORY.md and topic files), decide which notes are reliable enough to become permanent rules in CLAUDE.md or .claude/rules/, and remove stale or duplicate notes. It can also save a correction so the agent stops repeating a mistake, and turn a recurring fix into a standalone reusable skill.
Great for
- Developers who use Claude Code on a long-running project and want its memory kept tidy
- Teams that want recurring coding conventions turned into enforced project rules
- People who keep correcting the same agent mistakes across sessions
Skip it if
- Projects where the agent's memory should not be edited, since promoting or pruning entries changes future agent behaviour and can introduce wrong rules if reviewed carelessly
- Users of GitHub Copilot, where the SKILL.md notes say promotion is manual only
Try asking
- “What have you saved about this project in memory? Find entries worth making permanent.”
- “Promote our pnpm-not-npm convention from memory into CLAUDE.md.”
- “Audit my memory files and remove stale or duplicate entries.”
- “Remember that we always test API error responses, not just happy paths.”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 1 original files, the MIT licence and an attribution note.
self-improving-agent.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip self-improving-agent.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip self-improving-agent.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Review each promotion candidate before approving it, because promoted entries become enforced instructions.
- Run memory-status periodically to check line counts against the 200-line limit that truncates MEMORY.md.
- Keep promoted rules specific and scoped with .claude/rules/ where possible to avoid bloating CLAUDE.md.
What's inside
- SKILL.md
Redistributed unmodified from alirezarezvani/claude-skills under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.