# Improve Skill

> Convert cc-plugin-eval findings into a concrete rewrite brief, then hand off to skill-creator for the edit pass. Use when the user already ran an evaluation and now wants a fix plan, after asking "what should I fix first" or "rewrite this using the findings".

- Skill: `ancplua/improve-skill` (Agent Skill)
- Install (CLI): `npx skillmds@latest add ancplua/improve-skill`
- Raw SKILL.md: https://api.skillmd.com/api/skills/ancplua/improve-skill/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: ANcpLua (https://skillmd.com/u/ancplua)
- Updated: 2026-09-10
- Page: https://skillmd.com/skills/ancplua/improve-skill

---

# Improve Skill

Use this skill after `cc-plugin-eval` has already produced findings for a local skill.

The runnable commands below invoke the shipped script through the plugin root, which works in any checkout: `node "${CLAUDE_PLUGIN_ROOT}/scripts/cc-plugin-eval.js" <command> <args>`. Inline mentions of `cc-plugin-eval <command>` are shorthand for that same invocation; the bare `cc-plugin-eval` alias only exists if you ran `npm link` in the plugin directory.

## What This Skill Does NOT Do

This skill does not perform LLM-driven rewrites. That is the job of the `skill-creator` skill (Anthropic's `skill-creator@claude-plugins-official`). cc-plugin-eval produces structured findings and a rewrite brief; skill-creator turns the brief into edits with subagent-driven evals and grading. Keep the two boundaries clear.

## Workflow

1. Run `cc-plugin-eval analyze <skill-path> --brief-out ./skill-brief.json`. The same payload is also produced by `cc-plugin-eval improve <skill-path> --brief-out ./skill-brief.json`, which writes the improvement brief as the top-level payload.
2. Read the improvement brief. Group work into required fixes (severity `error` or `fail`) versus recommended fixes (severity `warn` or `info`).
3. Hand the brief to `skill-creator`. That skill specializes in skill rewrites with subagent-driven evals.
4. Re-run `cc-plugin-eval analyze <skill-path> --output ./after.json` and use `cc-plugin-eval compare ./before.json ./after.json --format markdown` to measure the delta.

## Focus Areas

- reduce trigger and invoke token costs (move bulky details into references, scripts, or helper files)
- keep `SKILL.md` compact (target under 500 lines; hard limit at 800; reference threshold is 350)
- improve trigger descriptions so Claude reliably routes to the skill (CC203, CC204)
- fix broken relative links (CC211), oversized descriptions (CC202), and frontmatter issues (CC213)
- normalize `allowed-tools` syntax (CC215) and add `disable-model-invocation` to task-style skills where appropriate (CC216)

## Chat Requests To Recognize

- `Improve this skill based on the evaluation.`
- `Rewrite this skill using the cc-plugin-eval findings.`
- `What should I fix first in this skill?`
- `Hand the brief to skill-creator.`

## Commands

```bash
node "${CLAUDE_PLUGIN_ROOT}/scripts/cc-plugin-eval.js" analyze <skill-path> --brief-out ./skill-brief.json
node "${CLAUDE_PLUGIN_ROOT}/scripts/cc-plugin-eval.js" improve <skill-path> --brief-out ./skill-brief.json --format markdown
node "${CLAUDE_PLUGIN_ROOT}/scripts/cc-plugin-eval.js" analyze <skill-path> --output ./after.json
node "${CLAUDE_PLUGIN_ROOT}/scripts/cc-plugin-eval.js" compare ./before.json ./after.json --format markdown
```

## Reference

- `../../references/chat-first-workflows.md`
- `../../references/evaluation-result-schema.md`