The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process project dependency manifests (e.g., package.json, requirements.txt, Pipfile) which are untrusted external inputs.
Ingestion points: The DependencyDiscovery class in resources/implementation-playbook.md identifies and reads various manifest and lock files from the project directory.
Boundary markers: No explicit delimiters or instructions are provided to the agent to treat the content of these manifests as untrusted or to ignore any embedded natural language instructions.
Capability inventory: The skill possesses capabilities that could be abused if an injection is successful, including reading arbitrary files in the codebase, performing network requests to external APIs, and generating automated remediation scripts or Pull Requests.
Sanitization: The provided code sketches do not include sanitization or validation of the manifest content before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill references and utilizes several external, well-known services and tools for its core functionality:
Fetches vulnerability advisory data from official registries including registry.npmjs.org, pypi.org, rubygems.org, and ossindex.sonatype.org.
Queries bundlephobia.com for package size analysis.
The provided GitHub Actions template includes commands to install third-party security tools such as safety via pip and license-checker via npx.
[COMMAND_EXECUTION]: The skill includes instructions and templates for executing shell commands within a CI/CD environment (GitHub Actions) to perform audits, check licenses, and generate reports.

codebase-cleanup-deps-audit