The Agent Skills Directory

[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it processes untrusted data from the user's project files.
Ingestion points: The script analyze_tests.py recursively searches and reads the contents of test files (e.g., .py, .js, .go, .java) and configuration files (e.g., conftest.py, package.json, pytest.ini) via the find_test_files and analyze_test_file functions.
Boundary markers: There are no markers or explicit instructions provided to the calling agent in SKILL.md to ignore instructions that might be embedded within the project files being analyzed.
Capability inventory: The skill is configured with Bash and Read tool access in SKILL.md, allowing it to run local scripts and read any file in the project.
Sanitization: The extraction logic in patterns.py (e.g., extract_pytest_fixtures) uses regular expressions to capture docstrings and fixture names directly from source code and includes them in the analysis summary without sanitization or escaping.

test-pattern-analysis