| name | find-awesome-skill |
| description | Use this skill when looking for curated skill recommendations from awesome lists, with exact install commands. |
| metadata | {"internal":true} |
Find Awesome Skill
Search the user's configured awesome skill sources, merge duplicates, and return concise recommendations with install commands.
Source discovery
Awesome sources are loaded from these layers, in this order:
.agents/awesome-skill-sources.local.json.agents/awesome-skill-sources.json~/.agents/awesome-skill-sources.json- The current repo's own
awesome-skills.json as the default source unless explicitly disabled
Duplicate sources are deduped by repo + path.
Use this section when the user asks which sources are searched, or when the CLI returns no results and you need to diagnose missing configuration.
Search flow
1. Primary path — use the CLI
Run the bundled finder with agent output (terse, LLM-readable):
npx cyberplace@<version> awesome find "<query>" --format agent
Read the output directly and present concise results with install commands. Default CLI output (without --format agent) is human-readable prose formatted for terminals.
2. Fallback — when the CLI cannot fully answer
Use manual reasoning from the source layers above when:
cyberplace is unavailable or awesome find fails- the query returns zero results — inspect configured sources and suggest
configure-awesome-sources
- the user asks which sources are searched (see Source discovery)
- you need to explain match rationale beyond JSON fields (corroboration across sources, tag vs summary nuance)
When falling back, merge duplicate repo or skill entries across sources and rank by:
- exact name match
- summary match
- tag match
- corroboration count
- source class preference
Return concise results with:
- repo or repo+skill identity
- why it matched
- primary
why_recommended
- whether other sources also recommended it
- exact
npx skills add ... install command
Guidance
- Exact name matches should be treated as the strongest signal.
- Tag and summary matches matter next.
- Corroboration across multiple awesome sources should increase confidence but should not hide why a result matched.
- If no results match, say so directly and suggest adding or configuring more awesome sources.
