Analyze and optimize a skill for discoverability on ClawHub, skills.sh, and similar skill directories. Use when you need to improve naming, slug choice, SKIL...
Check that the skill remains broadly usable in OpenClaw, Claude Code, Codex, and Cursor without adding platform-specific optimization unless the user asks for it.
Rewrite only what improves discovery or trigger quality. Preserve the skill's actual capability boundaries.
If the user asks for implementation, edit the target skill and re-run the analyzer to confirm improvement.
Definition of Done
The analyzer has run and a baseline report exists.
All high severity findings have been addressed or explicitly justified.
The description covers at least 3 query classes: exact task phrase, synonyms, and user-intent language.
Example prompts in SKILL.md reflect realistic user phrasing.
Target platform compatibility status is explicit: verified where evidence exists, otherwise unverified.
Re-running the analyzer shows improvement over the baseline.
Output Requirements
Produce a concise report with:
Current strengths and weaknesses
Platform-specific findings for ClawHub and skills.sh
Compatibility notes for OpenClaw, Claude Code, Codex, and Cursor using verified / unverified language only when relevant
Missing keywords, synonyms, and user-phrased queries
Recommended rewrites for name, slug, description, and first-screen content
Trust and conversion gaps such as missing prerequisites, examples, screenshots, badges, stars, or install guidance
A prioritized action list with high, medium, and low impact items
Optimization Heuristics
Prefer concrete, searchable names over abstract brand names.
Put the problem, object, and action into the first 1 to 2 sentences.
Include realistic user phrasings and close synonyms in description and the top of SKILL.md.
Make the skill's boundary explicit so retrieval stays precise.
Keep frontmatter concise but query-dense.
Put examples near the top; examples improve both semantic recall and click-through.
Add references only when they deepen a repeated workflow or expose non-obvious domain language.
Avoid generic filler such as "powerful", "seamless", or "all-in-one" unless backed by specifics.
Treat eval.yaml and UI metadata as optional enhancements, not baseline listing requirements.
Rewrite Formula
Use this pattern when rewriting a weak listing:
Then reinforce it with:
3 to 5 example prompts using realistic user language
A first-screen summary that repeats the main task phrase once
One or two trust signals such as prerequisites, constraints, or costs
Competitor Comparison
When a user asks for a market-aware rewrite:
Identify 2 or 3 comparable skills on the same directory.
Compare names, descriptions, example prompts, and first-screen structure.
Note what the target skill is missing, not just what competitors include.
Recommend changes that improve clarity and trust without copying competitors' wording.
Anti-Patterns
Do not do any of the following:
Suggest fake installs, fake stars, fake usage, or other manipulative ranking tactics
Recommend version bumps or metadata churn without a real content change
State platform behavior as a guaranteed fact when it is only an observed heuristic
Stuff unrelated keywords into the description or examples
Broaden a skill's positioning beyond what the implementation can actually do
Flag missing eval.yaml or agents/openai.yaml as if they were mandatory for ClawHub or skills.sh publication
When Editing a Skill
Preserve the existing capability surface unless the user asks to expand it.
Prefer improving description, top-of-file wording, and usage examples before adding large new references.
If the folder name is weak for search, recommend a slug rename explicitly instead of renaming silently.
If the target platform has already indexed the old name, explain the migration tradeoff.