The Agent Skills Directory

[SAFE]: Legitimate Tooling: The skill utilizes standard project commands like make test, git diff, and uv run pytest to achieve its goal. These actions are consistent with the primary purpose of the skill in a software development context.
[PROMPT_INJECTION]: Indirect Prompt Injection Surface (Category 8): (1) Ingestion points: The skill analyzes the output generated by make and pytest. (2) Boundary markers: The instructions do not specify delimiters to separate untrusted test output from the agent's instructions. (3) Capability inventory: The agent has the ability to execute shell commands and modify code. (4) Sanitization: No sanitization of test output is described. This represents an inherent surface area for tools processing logs but does not indicate malicious intent.

test-fixing