The Agent Skills Directory

[SAFE]: The skill's primary function is to read project metadata (like README.md, package.json, and requirements.txt) and write documentation files (AGENTS.md, CLAUDE.md). This is a standard administrative task for maintaining AI-assisted development environments.
[SAFE]: There are no detected patterns of prompt injection, data exfiltration, or unauthorized command execution. The 'Operational Commands' mentioned (e.g., 'bun run dev') are instructions for the agent to include as text in documentation, not commands to be executed by the skill itself.
[SAFE]: The skill uses direct instructions and autonomous file writing ('Overwrite Authority') to automate the setup of governance files. While this reduces user oversight during the generation process, the resulting files are standard markdown and do not introduce system-level vulnerabilities.
[SAFE]: The skill exhibits an Indirect Prompt Injection surface because it ingests untrusted data from the local project (e.g., README.md) to generate instructions. However, since the output is restricted to documentation files and the skill does not possess high-privilege execution capabilities, the risk is negligible.
Ingestion points: README.md, docs/, CONTRIBUTING.md, package.json, requirements.txt, Cargo.toml.
Boundary markers: None present.
Capability inventory: File creation and modification (Markdown files).
Sanitization: None.

agents-md