The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local shell commands and vendor-provided scripts to audit the repository environment.
Evidence: Uses standard tools such as gh, grep, find, jq, and git for repository analysis in SKILL.md and scripts/repo-health-check.workflow.js.
Evidence: Executes maintenance scripts located within the vendor's directory, specifically daymade-claude-code/marketplace-dev/scripts/check_marketplace.sh and check_doc_skill_lists.py.- [PROMPT_INJECTION]: The skill processes potentially untrusted content from GitHub Pull Requests and Issues, creating a surface for indirect injection.
Ingestion points: Fetches data via gh pr list, gh issue list, gh pr view, and gh issue view in scripts/repo-health-check.workflow.js.
Boundary markers: Agents are constrained by dimension-specific prompts and a strict JSON schema (CHECK_SCHEMA) for their output.
Capability inventory: The agents have access to shell utilities (grep, find) and the repository's files.
Sanitization: The skill incorporates a robust "Counter-Review" methodology as described in SKILL.md and references/health-check-methodology.md, requiring the agent to independently verify and validate findings before reporting them to the user to prevent the relaying of noise or malicious instructions embedded in the analyzed data.

marketplace-health-check