Skip to content
Agentmart
Submit a tool

self-improving

Capture learnings, errors, and corrections from conversations for continuous self-improvement.

Skill by zzycxz

IndexedMIT★ 9
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/zzycxz-self-improving

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Capture learnings, errors, and corrections from conversations for continuous self-improvement. Use when: completing a complex task, encountering an error and fixing it, learning a user preference, or discovering a better approach.

Use cases: productivity

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/self-improving

What's inside

Skill name
self-improving
Description
Capture learnings, errors, and corrections from conversations for continuous self-improvement. Use when: completing a complex task, encountering an error and fixing it, learning a user preference, or discovering a better approach.
License
MIT
Instructions
40 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Self-Improving Agent

Automatically capture insights from task execution to improve future performance.

## When to Record Learnings

Record a learning when:

1. **Error → Fix pattern**: You encountered an error and found a working solution
2. **User correction**: The user corrected your approach or output
3. **Preference discovery**: You learned the user prefers a specific style, format, or tool
4. **Optimization found**: You discovered a faster or better way to accomplish something
5. **Tool discovery**: You found a tool combination that works well for a specific domain

## How to Record

Use the `memory_add_fact` tool to store learnings as structured facts:

```
Type: learning
Content: "When doing X, prefer Y because Z"
Confidence: high|medium|low
```

## Categories

- **tool-pattern**: Effective tool combinations or sequences
- **error-recipe**: Common errors and their fixes
- **user-preference**: User-specific preferences (style, format, language)
- **performance**: Faster approaches or shortcuts discovered
- **domain-knowledge**: Facts learned about the user's project/domain

## Anti-Patterns (Do NOT Record)

- Obvious or trivial facts
- Temporary state (file paths, current date)
- Speculative guesses without evidence
- Duplicate or contradictory existing learnings

From zzycxz/rooster, MIT.

More from zzycxz/rooster

Signals

Stars
9
Forks
0
Open issues
0
Last commit (this folder)
Jun 9, 2026 (4 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from zzycxz

Cite this page

self-improving skill. Agent Mart. https://agentmart.dev/skills/zzycxz-self-improving (last updated 2026-10-24)