The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for running common developer tools such as git, npm audit, pip-audit, and govulncheck. These commands are used to identify tech debt and security vulnerabilities within the project context.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted repository data (source code, dependency files) to generate a health report. Ingestion points include project files such as package.json and .env (SKILL.md). No specific boundary markers or sanitization logic is defined for the tool outputs. Capability inventory includes file system access and package managers. While malicious content in audited files could theoretically influence the agent's summary, the risk is inherent to any code auditing tool.
[DYNAMIC_EXECUTION]: The skill utilizes npx to execute utilities like depcheck and ts-prune. These are standard Node.js development tools and are used here in a controlled diagnostic capacity.

repo-hygiene