Supply chain intelligence and ecosystem analysis for OpenClaw skills. Operates offline-first against locally installed skills; optionally enriches with live clawhub metadata (stars, downloads, moderation verdicts) when network is available.
| Trigger | Action |
|---|---|
| --------- | -------- |
| "Analyze my skills" / "skill inventory" | ingest analyze-all (one-shot) |
| "What tools / packages do my skills use?" | analyze packages |
| "Show dependencies for skill X" | analyze supply-chain --skill |
| "Which skills share package X?" | analyze find-users --package |
| "Skill categories / ecosystem breakdown" | analyze categories |
| "How complete / healthy are my skills?" | analyze health |
| "Do any skills overlap or duplicate?" | analyze overlaps |
| "Sync online popularity / security data" | ingest enrich |
| "Full ecosystem report with insights" | analyze report |
| "Rebuild graph from scratch" | ingest reset && ingest scan |
# Reset, scan, health check, overlap analysis, full report — all in one command
python3 scripts/ingest.py analyze-all
# With custom directories
python3 scripts/ingest.py analyze-all --dirs ~/.cursor/skills-cursor ~/.openclaw/skills ~/Downloads/skills-main/skills
# Auto-discover skills under default paths
python3 scripts/ingest.py scan
# Specify directories explicitly
python3 scripts/ingest.py scan --dirs ~/Downloads/skills-main/skills ~/.codex/skills ~/Downloads/ontology
# Online enrichment: adds stars, downloads, moderation verdict from clawhub
# Skipped automatically if network is unavailable
python3 scripts/ingest.py enrich
# Check what's currently in the graph
python3 scripts/ingest.py status
# Reset and rebuild
python3 scripts/ingest.py reset
python3 scripts/ingest.py scan
# Ecosystem overview (counts, categories, top packages)
python3 scripts/analyze.py stats
# Category distribution
python3 scripts/analyze.py categories
# Skill completeness health scores (0-100 per skill, with specific issues)
python3 scripts/analyze.py health
# Detect overlapping or complementary skill pairs
python3 scripts/analyze.py overlaps
# Top N skills by a metric
python3 scripts/analyze.py top --by stars --limit 10
python3 scripts/analyze.py top --by downloads --limit 10
# Skill profile card (dependencies, tools, bins, invocation pattern, online metrics)
python3 scripts/analyze.py profile --skill ontology
# Full supply chain for a skill (invoked_via → requires_bin → tools → packages → skill deps)
python3 scripts/analyze.py supply-chain --skill agent-browser
# Find all skills that use a specific package
python3 scripts/analyze.py find-users --package playwright
# Most-used packages across all skills
python3 scripts/analyze.py packages --top 20
# Full markdown report with Key Insights section
python3 scripts/analyze.py report
| Source | What it provides | Required? |
|---|---|---|
| -------- | ----------------- | ----------- |
SKILL.md frontmatter | name, description, license, allowed-tools, metadata.requires.bins, read_when | Yes |
requirements.txt | Declared Python package deps (pypi) | When present |
pyproject.toml | Python deps: PEP 621, Poetry, optional-deps | When present |
Pipfile | Python deps (Pipfile format) | When present |
package.json | npm deps: dependencies / devDeps / peerDeps | When present |
scripts/*.py AST scan | Implicit Python imports (non-stdlib only) | When present |
_meta.json / .clawhub/origin.json | slug, version, registry | When present |
| clawhub API (online) | stars, downloads, moderation verdict, owner | Optional |
| Relation | Meaning |
|---|---|
| ---------- | --------- |
belongs_to_category | skill belongs to a functional category |
requires_package | skill depends on a pypi / npm package |
uses_tool | skill uses a detected tool (heuristic) |
requires_tool | skill requires a system binary (metadata.requires.bins) |
invoked_via | skill is called through a tool channel (allowed-tools) |
depends_on_skill | skill references another skill in its description |
The ingest script checks these locations by default:
~/.openclaw/skills~/.openclaw/extensions (extensions exposing their own skills via /skills/* /Applications/OpenClaw.app/Contents/Resources/skills (macOS app bundle, when installed)In addition, ingest scan augments the user-provided or default --dirs
arguments with:
/skills inside a repository
$(npm root -g)/openclaw/skills — globally installed OpenClaw skills froma Node.js / npm environment
Any directory containing a SKILL.md file is treated as a skill.
ontology:
reads: [Skill, SkillCategory, Tool, SoftwarePackage]
writes: [Skill, SkillCategory, Tool, SoftwarePackage]
storage: memory/skillchain/graph.jsonl
schema: schema/skillchain.yaml
preconditions:
- "At least one local skill directory is accessible"
postconditions:
- "Every discovered skill has a Skill entity"
- "Skills with requirements.txt have SoftwarePackage entities and requires_package relations"
Graph data is written to memory/skillchain/graph.jsonl using the same append-only JSONL format as the ontology skill. Reuse scripts/ontology.py from the ontology skill for low-level graph operations.
references/model.md — Full ontology model (types, relations, constraints)schema/skillchain.yaml — Machine-readable schema for validation共 1 个版本