The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a local Python script, scripts/docs-inventory.py, which uses subprocess.run to invoke the git log command. This is used strictly to retrieve file modification timestamps for documentation inventory purposes. The implementation uses a list-based argument format and does not use a shell, minimizing the risk of command injection from file paths.
[INDIRECT_PROMPT_INJECTION]: As a documentation analysis tool, the skill is designed to ingest and process untrusted external documentation files (.md, .mdx, .rst) from user-specified directories. While a malicious file could contain instructions intended to influence the agent's audit findings, this is an inherent risk of the skill's primary function.
Ingestion points: Documentation files are read by both the inventory script and the agent during the manual audit phase.
Boundary markers: None are explicitly defined in the instructions for separating documentation content from audit instructions.
Capability inventory: The agent can read files, execute the included inventory script, and output reports based on its findings.
Sanitization: The skill does not implement specific sanitization or escaping of documentation content before the agent evaluates it for classification.

developer-docs-structure-audit