The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill reads and processes external Python source files, which is a known vector for indirect prompt injection.
Ingestion points: Python files (.py) and directories processed via ast or inspect modules.
Boundary markers: Absent. The instructions do not specify any delimiters or warnings to the agent to ignore instructions embedded within the extracted docstrings.
Capability inventory: File system read (to access source code) and typically file system write (to save generated markdown), though write is not explicitly defined in the script.
Sanitization: None. The skill suggests direct extraction and formatting of docstring content without sanitization.
[Dynamic Execution] (LOW): The documentation suggests using the inspect module for runtime introspection.
Evidence: Using inspect often requires importing the target module into the Python environment. Importing untrusted code can trigger execution of top-level logic or malicious code within the module's __init__ or global scope.

doc-generator