moonrunnerkc/skillcheck
Cross-agent skill quality gate for SKILL.md files. Validates against the agentskills.io specification.
View on GitHubTrust Signals
- Scorecard Score
- not yet scored
- Maintenance Recency
- Activelast commit Jul 9, 2026
- License
- MIT
Pinned Snippet
uses: moonrunnerkc/skillcheck@f2c2946e7007cc745a66d365849a1a740b24fb2c # v1.4.1tags can be moved; commit SHAs can't. why a SHA?
Inputs
| name | description | required | default |
|---|---|---|---|
| path | Path to a SKILL.md file or directory to scan recursively. | no | . |
| version | skillcheck version to install (e.g., '1.2.3'). Leave empty for latest. | no | "" |
| min-desc-score | Minimum description quality score (0-100). Below this triggers a warning. | no | "" |
| target-agent | Scope compat checks: claude, vscode, cursor, or all (default: all). | no | "" |
| strict-vscode | Promote VS Code compat issues to errors. | no | false |
| strict-cursor | Promote Cursor compat issues to errors. | no | false |
| strict | Strict mode. Escalates warnings to exit 1 and enables --strict-vscode and --strict-cursor. | no | false |
| skip-dirname-check | Skip directory-name matching check. | no | false |
| skip-ref-check | Skip file reference validation. | no | false |
| ignore | Comma-separated rule prefixes to suppress (e.g., 'sizing,disclosure'). | no | "" |
| max-lines | Override the line-count threshold (default: 500). | no | "" |
| max-tokens | Override the token-count threshold (default: 8000). | no | "" |
| format | Output format. Defaults to github for automatic PR annotations. | no | github |
| semantic | Enable semantic-adjacent validation. In standalone CI this enables heuristic graph analysis. | no | false |
| analyze-graph | Run heuristic capability graph analyzers and merge diagnostics into the report. | no | false |
| ingest-critique | Path to an agent self-critique JSON response to ingest. | no | "" |
| critique-agent | Critique prompt/source agent: claude, codex, or cursor. | no | "" |
| ingest-graph | Path to an agent graph-extraction JSON response to ingest. | no | "" |
| graph-agent | Graph prompt/source agent: claude, codex, or cursor. | no | "" |
| history | Append a validation record to .skillcheck-history.json. | no | false |
| fail-on-regression | With --history, exit 1 on history.skill.regressed. Independent of --strict. | no | false |
| explain-score | Show per-dimension breakdown for description quality scores. | no | false |
| activation-hypotheses | Generate activation hypotheses. This is an emit mode and does not produce PR annotations. | no | false |
| tiktoken | Install the optional tiktoken extra for more accurate token estimates. | no | false |
Outputs
no outputs