The Agent Skills Directory

[PROMPT_INJECTION]: The generate-docs.py script contains a vulnerability surface for indirect prompt injection by processing untrusted data.
Ingestion points: The script reads file content from a path provided via command-line arguments (sys.argv[1]).
Boundary markers: There are no delimiters or specific instructions to the agent to ignore embedded instructions within the extracted docstrings.
Capability inventory: The script performs local file reading and outputs text to stdout; the agent is intended to use this output to create documentation files.
Sanitization: The script extracts and prints docstrings and parameter names directly from the source code using the ast module without sanitization or escaping.
[SAFE]: The skill refers to official documentation from Anthropic at https://code.claude.com/docs/en/skills.

doc-generator