The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes Python scripts (create_test_session.py and verify_test_session.py) that perform local file system operations to initialize and validate test sessions. These scripts are safe because they implement strict path validation using Path.resolve() and relative_to to ensure all operations remain within the project's root directory, effectively preventing directory traversal and symlink-based attacks.
[PROMPT_INJECTION]: The skill instructions utilize directives like 'Independent Review' and 'Ignore previous plans' as functional requirements for the testing methodology. These are designed to reduce cognitive bias during QA and do not constitute attempts to bypass AI safety guardrails or override core agent instructions.
[DATA_EXFILTRATION]: No network activity or unauthorized data transmission patterns were detected. Security-related strings such as API keys or system paths found in the documentation are provided solely as educational examples of vulnerabilities to be tested for in other projects.
[REMOTE_CODE_EXECUTION]: The skill does not perform any remote downloads or dynamic execution of untrusted external code. All script logic is local and uses predefined templates for generation.

auto-test-skill