The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/doc_audit.py executes the local git binary using subprocess.run to determine the project root and retrieve file modification timestamps. This is implemented securely with static arguments and does not invoke a shell environment.
[PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection by scanning external documentation and codebase files (ingestion points). While the subagents have the capability to write to the file system (capability inventory), the prompt templates currently lack explicit boundary markers or content sanitization. This is documented as a standard risk factor for tools that process untrusted text, but no malicious intent was found.

doc-maintenance