Skip to content

cli-reference.md

Latest commit

 

History

History
88 lines (66 loc) · 3.82 KB

cli-reference.md

File metadata and controls

88 lines (66 loc) · 3.82 KB

CLI Reference

All commands are exposed as npm scripts and run through src/cli.ts.

PlugScout scope order is: Claude plugins, Claude connectors, Copilot extensions, Skills, MCP servers.

Packaged CLI behavior:

  • plugscout with no arguments opens the home screen.
  • plugscout upgrade check performs a release check against GitHub Releases.
  • plugscout web --open writes a readable browser report and opens it.
  • --no-update-check disables the daily automatic update check for the current command run.

Core Operations

  • npm run about
  • npm run status [-- --verbose]
  • npm run sync [-- --kind skill,mcp,claude-plugin,claude-connector,copilot-extension] [-- --dry-run]

Guided Setup and Diagnostics

  • npm run setup — one-step: install prerequisites + write default config + sync all catalogs
  • npm run init [-- --project .] — interactive config init only
  • npm run doctor [-- --project .]
  • npm run doctor -- --project . --install-deps
  • setup is the recommended first command after installing the package; it calls init non-interactively and runs an initial sync
  • init default kinds include skill,mcp,claude-plugin,claude-connector,copilot-extension
  • riskPosture=strict makes list and recommend default to safe-only views
  • doctor --install-deps bootstraps the supported skills CLI when it is missing
  • supported legacy MCP installs now prefer direct npm install -g or docker pull when the target is unambiguous

Catalog Discovery and Inspection

  • npm run list -- --kind mcp --limit 10
  • npm run list -- --kind claude-plugin --limit 20 (includes source, catalog type, confidence columns)
  • npm run list -- --kind claude-connector --limit 20 (separate Claude Connectors catalog)
  • npm run list -- --kind claude-plugin --limit 20 --readable (wrapped wide table mode)
  • npm run list -- --kind claude-plugin --limit 10 --details (per-item decision evidence: trust/risk/provenance/install)
  • npm run search -- security
  • npm run explain -- --limit 20
  • npm run explain -- --kind mcp --format json
  • npm run scan -- --project . --format table
  • npm run scan -- --project . --format json --out scan-report.json
  • npm run show -- --id mcp:filesystem
  • npm run show -- --id claude-connector:asana (prints provenance source/sourcePage when present)
  • npm run top -- --project . --limit 5
  • npm run top -- --project . --limit 5 --details (score equation + block reason + install hint per item)

Recommendation and Export

  • npm run recommend -- --project . --format table
  • npm run recommend -- --project . --only-safe --sort trust --limit 10 --explain-scan
  • npm run recommend -- --project . --only-safe --sort trust --limit 10 --details
  • npm run recommend -- --project . --llm --explain-scan
  • npm run recommend -- --project . --export csv --out recommendations.csv
  • npm run recommend -- --project . --export md --out recommendations.md

Risk and Installation Controls

  • npm run assess -- --id mcp:remote-browser
  • npm run install:item -- --id mcp:filesystem --yes --install-deps
  • npm run install:item -- --id mcp:remote-browser --yes --override-risk
  • npm run install:item -- --id skill:playwright --yes --install-deps

Security Operations

  • npm run whitelist:verify
  • npm run quarantine:apply -- --report data/security-reports/YYYY-MM-DD/report.json

Version and Upgrade Checks

  • plugscout upgrade check
  • plugscout recommend --project . --no-update-check
  • plugscout status --no-update-check

Web Report

  • plugscout web --out .plugscout/report.html
  • plugscout web --kind claude-connector --limit 200 --open (includes score legend + decision cards per item)

Legacy Ingestion and Validation

  • npm run validate:data
  • npm run ingest:skills
  • npm run ingest:mcps
  • npm run ingest:claude-plugins
  • npm run ingest:copilot-extensions
  • npm run ingest:all