Improve existing agent skills based on user feedback and best practices. Use when the user wants to fix, enhance, or refactor an existing skill. Gathers user feedback first, then applies technical analysis and implements improvements.
This skill inherits all available tools. When active, it can use any tool Claude has access to.
Improve existing agent skills through gathering user feedback and technical analysis.
Ask user for the skill path if not provided (e.g., .claude/skills/skill-name/).
Read SKILL.md and understand current structure.
Ask the essential question:
"What problems or improvements do you want for this skill?"
Based on the response, ask follow-up questions as needed:
| If user mentions... | Follow-up question |
|---|---|
| Vague dissatisfaction | "Can you give a specific example of when it didn't work well?" |
| Trigger issues | "When does it fail to trigger, or trigger unexpectedly?" |
| Missing features | "What should the output look like when this feature works?" |
| Too much/little output | "What level of detail do you prefer?" |
| Unclear scope | "What tasks should this skill handle vs. not handle?" |
Keep follow-ups minimal. One or two is usually enough.
Check against best practices (reference reviewing-skills criteria):
Organize improvements into:
User-requested improvements:
Technical improvements:
Format:
## Improvement Plan
### User-requested
1. [Issue from feedback] → [Proposed fix]
### Technical
1. [Issue from analysis] → [Proposed fix]
Proceed with these changes?
Execute the improvements:
Show diff or summary of changes made.
Quick check:
| reviewing-skills | improving-skills |
|---|---|
| Technical analysis only | User feedback + technical analysis |
| Provides feedback | Implements changes |
| Read-only | Read-write |
{
"permissions": {
"allow": [
"Skill(improving-skills)"
]
}
}