Search Skill
Discover existing skills and plugins from curated sources, evaluate quality with toothbrush filter, output comparison table. Plugins bundle skills/commands/agents/MCP — mine them for either. Philosophy: inspect, adapt, own — never import blindly 🪥
Workflow
- Parse query from
$ARGUMENTS(e.g., "pdf converter", "git workflow") - Search tiered sources (parallel where possible):
| Tier | Source | Method |
|---|---|---|
| 1 | Anthropic skills | WebSearch: "$QUERY claude code skill site:github.com/anthropics" |
| 1 | Anthropic plugins (official) | WebFetch raw .claude-plugin/marketplace.json from anthropics/claude-plugins-official, grep $QUERY |
| 1 | VoltAgent curated | WebSearch: "$QUERY site:github.com/VoltAgent/awesome-agent-skills" |
| 2 | Anthropic plugins (community) | WebFetch raw .claude-plugin/marketplace.json from anthropics/claude-plugins-community (security-screened), grep $QUERY |
| 2 | GitHub community | WebSearch: "$QUERY claude code SKILL.md site:github.com" |
| 2 | Awesome lists | WebSearch: "$QUERY awesome-claude-skills OR awesome-claude-code" |
| 3 | claudemarketplaces / aiskillstore | WebSearch: "$QUERY site:claudemarketplaces.com" · aggregators |
| 3 | SkillHub / broad | WebFetch: skillhub.club · WebSearch: "$QUERY claude code skill 2026" |
- Fetch SKILL.md for top candidates via
WebFetchon raw GitHub URLs - Apply toothbrush filter — see
reference.mdfor rubric details:- Token efficiency: <500 ideal, hard fail >2000
- Clear triggers: keywords + file types required
- Single responsibility: one capability only
- Anthropic patterns + progressive disclosure (soft)
- Output comparison report — see
reference.md § Output Report Template(table: Skill · Source · ~Tokens · Quality · Verdict, then Recommended approach + Key patterns)
Rules
- Output verdicts (study/adapt/ignore), NOT install commands
- Skip tiers 2-3 if tier 1 has strong matches
- Fetch actual SKILL.md source when possible — evaluate code not marketing
- See
reference.mdfor full source registry and quality rubric weights