The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/docs-inventory.py script executes local shell commands via the subprocess.run function to query git for the last modification date of documentation files. This is a legitimate operation within the context of a documentation audit and uses secure argument passing to minimize injection risks.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external documentation content from user-specified repositories or live websites. This creates an attack surface where malicious instructions embedded in the documentation could attempt to influence the agent's analysis or subsequent actions.
Ingestion points: The agent reads documentation files (Markdown, reStructuredText) and navigation configurations (YAML, JSON, etc.) from the user's project directory.
Boundary markers: There are no explicit instructions to the agent to disregard instructions found within the processed documentation files.
Capability inventory: The skill allows for local file reading and script execution (scripts/docs-inventory.py).
Sanitization: The skill analyzes content for structural patterns (headings, code blocks, etc.) but does not explicitly sanitize the text before the agent processes it for qualitative judgment.

developer-docs-structure-audit