The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes three internal Python scripts (coverage-analyzer.py, edge-case-generator.py, and flaky-test-analyzer.py) using the uv run command. These scripts perform localized data processing of coverage reports, function signatures, and test logs. All execution is limited to the skill's own internal logic and does not involve external code.
[INDIRECT_PROMPT_INJECTION]: The skill analyzes user-provided source code and test logs to identify coverage gaps and edge cases. This creates an input ingestion surface where external data enters the agent context. However, the risk is mitigated by the skill's focus on structural and type-based analysis rather than natural language interpretation of the input code.
[SAFE]: The skill includes strong safety guidelines that prevent it from running tests or modifying files. It operates entirely on local project data and does not perform any network operations or external downloads. All scripts use standard parsing techniques without unsafe dynamic execution (e.g., eval/exec).

test-architect