Claude Skill

skill-evolve

A meta-skill that takes feedback after a skill runs and automatically improves the corresponding SKILL.md. Use when a skill was wrong or incomplete, or to record a newly discovered gotcha.

LLM Mart · 0 points · 14 views 0 listing impressions 0 install-command copies
Virus-scanned Reviewed automatically before listing.

Full trust report

Download leeyudok-agents-scaffold-presets_lang-en_base_.claude_skills_skill-evolve-1b2f034.zip · 1 KB
Part of leeyudok/agents-scaffold — 32 skills

Install

skills CLI npx skills add https://github.com/LeeYudok/agents-scaffold/tree/main/presets/lang-en/base/.claude/skills/skill-evolve
Claude Code claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install leeyudok-agents-scaffold@llmmart
Git git clone https://github.com/LeeYudok/agents-scaffold.git

The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole leeyudok/agents-scaffold collection as a plugin from our marketplace. Git is the plain clone.

Skill manifest

Skill Self-Improvement

Origin: the skill of the same name in leeyudok/doksam-skills. Deliberately forked for template bundling — not auto-synced; good improvements are cherry-picked manually.

Reflects problems found during execution and user feedback into the target skill's SKILL.md.

Trigger

/skill-evolve <skill-name> <feedback>

Process

  1. Read the target: Read the full .claude/skills/<skill-name>/SKILL.md.
  2. Analyze: Analyze the feedback + recent execution context → decide what to change.
    • New gotcha → add to the ## Learned warnings section with a date
    • Step correction → update the process section
    • Wrong command → fix it
  3. Propose: Quote the existing content + present the change as a diff (added/removed lines).
  4. Ask for confirmation: "Apply this change? (Y/n)"
  5. Apply: After confirmation, edit SKILL.md directly.
  6. Commit: After re-confirming the branch:
    git add .claude/skills/<skill-name>/SKILL.md
    git commit -m "evolve skill/<skill-name>: <summary>"
    
  7. Verify (mini eval): Run the modified skill on the same kind of input that triggered the improvement, and confirm via a before/after comparison that the previously failing part is actually fixed.

Rules

  • Never delete existing Learned warnings — only accumulate
  • Date tag required: (YYYY-MM-DD)
  • Merge duplicate warnings
  • Re-confirm the branch right before committing (avoid committing on main)

Output format

Proposed improvement
Existing: [quote]
Proposed: [revision]

diff:
+ added line
- removed line

Apply this change? (Y/n)

Learned warnings

(Notes discovered during execution accumulate here)

Files (agents-scaffold)
  • SKILL.md 2 KB
    ---
    name: skill-evolve
    description: A meta-skill that takes feedback after a skill runs and automatically improves the corresponding SKILL.md. Use when a skill was wrong or incomplete, or to record a newly discovered gotcha.
    argument-hint: "[name of skill to improve] [feedback/improvement]"
    allowed-tools: Edit, Read, Write, Bash
    ---
    
    # Skill Self-Improvement
    
    > Origin: the skill of the same name in [leeyudok/doksam-skills](https://github.com/leeyudok/doksam-skills).
    > **Deliberately forked** for template bundling — not auto-synced; good improvements are cherry-picked manually.
    
    Reflects problems found during execution and user feedback into the target skill's SKILL.md.
    
    ## Trigger
    
    `/skill-evolve <skill-name> <feedback>`
    
    ## Process
    
    1. **Read the target**: Read the full `.claude/skills/<skill-name>/SKILL.md`.
    2. **Analyze**: Analyze the feedback + recent execution context → decide what to change.
       - New gotcha → add to the `## Learned warnings` section with a date
       - Step correction → update the process section
       - Wrong command → fix it
    3. **Propose**: Quote the existing content + present the change as a diff (added/removed lines).
    4. **Ask for confirmation**: "Apply this change? (Y/n)"
    5. **Apply**: After confirmation, edit SKILL.md directly.
    6. **Commit**: After re-confirming the branch:
       ```bash
       git add .claude/skills/<skill-name>/SKILL.md
       git commit -m "evolve skill/<skill-name>: <summary>"
       ```
    7. **Verify (mini eval)**: Run the modified skill on the same kind of input that triggered the improvement, and confirm via a before/after comparison that the previously failing part is actually fixed.
    
    ## Rules
    
    - Never delete existing Learned warnings — only accumulate
    - Date tag required: `(YYYY-MM-DD)`
    - Merge duplicate warnings
    - Re-confirm the branch right before committing (avoid committing on main)
    
    ## Output format
    
    ```
    Proposed improvement
    Existing: [quote]
    Proposed: [revision]
    
    diff:
    + added line
    - removed line
    
    Apply this change? (Y/n)
    ```
    
    ## Learned warnings
    
    (Notes discovered during execution accumulate here)
    

Comments (0)

Sign in to join the conversation.

No comments yet.

Reviews (0)

No reviews yet.

Related