The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions involve executing pytest to gather test results and coverage data. This capability executes any Python code located in the project's test directory.
[PROMPT_INJECTION]: The skill extracts docstrings and comments from test files and includes them in the generated documentation, creating a vulnerability to indirect prompt injection where malicious instructions in a test file could influence the agent's behavior.
Ingestion points: Python source files and test suites processed by scripts/extract_tests.py and read via cat or grep.
Boundary markers: Absent; the extracted text is directly inserted into markdown templates without delimiters or 'ignore' instructions.
Capability inventory: The skill utilizes find, cat, grep, and pytest, and executes an internal Python script to analyze code.
Sanitization: Absent; the script extracts docstrings and code comments without sanitizing or validating the content for potential injection patterns.

test-case-documentation