| name | source-discovery |
| description | This skill should be used when the user asks to "find sources", "discover URLs", "catalog a source", "classify a URL", "search for libraries", "check what exists", or mentions source discovery, URL cataloging, source classification, or research source management.
|
Source Discovery
Source Discovery Protocol
When discovering sources, follow this workflow:
- Search — Check existing source-catalog.md for duplicates
- Fetch — Use
npx agent-fetch "<url>" --json for full content
- Classify — Assign priority: HIGH / MEDIUM / LOW / SKIP
- Catalog — Append to
docs/research/source-catalog.md
- Record — Write YAML provenance for HIGH/MEDIUM sources
Classification Criteria
| Priority | Criteria |
|---|
| HIGH | Directly addresses current research question, authoritative source |
| MEDIUM | Related context, community discussion, alternative approach |
| LOW | Tangentially related, outdated but historically relevant |
| SKIP | Irrelevant, broken link, or duplicate of existing source |
Source Catalog Format
Append entries to docs/research/source-catalog.md:
| [status] | Source Name | URL | Type | Verdict |
Rules
- Log EVERY URL encountered — never silently discard
- Use
npx agent-fetch for content, never WebFetch (truncates)
- Check for duplicates before adding to catalog
- HIGH sources get immediate YAML provenance records
- Include verdict explaining why the source was classified at that level