The Agent Skills Directory

[COMMAND_EXECUTION]: The skill employs standard shell utilities such as mv, cp, ln, and test to perform file management tasks, including backing up documentation and creating symlinks (e.g., linking CLAUDE.md to AGENTS.md).
[DATA_EXFILTRATION]: The skill inspects project configuration files like package.json, pyproject.toml, and .env.example to extract metadata for documentation. This is a local-only operation with no evidence of network transmission or data exfiltration patterns.
[INDIRECT_PROMPT_INJECTION]: The skill processes existing documentation and configuration files which could contain instructions intended to influence the agent's behavior during the update process.
Ingestion points: Reads contents from AGENTS.md, README.md, CONTRIBUTING.md, and various package manager configuration files.
Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions when reading these files.
Capability inventory: Includes file system read/write, directory listing (ls, tree), and file manipulation (mv, cp, ln).
Sanitization: No explicit sanitization or validation of the ingested documentation content is performed before processing.

md-docs