The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection by design, as it ingests untrusted data from the local filesystem to generate its outputs.
Ingestion points: In SKILL.md, the agent is instructed to inspect pyproject.toml, noxfile.py, .config/mise/, and existing test files (tests/conftest.py).
Boundary markers: No explicit delimiters or instructions to ignore embedded malicious content are present in the skill instructions.
Capability inventory: The skill utilizes command execution capabilities via uv run pytest, uv run nox, and mise run, which could be influenced by injected content in the configuration files.
Sanitization: There is no evidence of sanitization or validation of the content read from these repository files before processing.
[COMMAND_EXECUTION]: The skill explicitly directs the agent to execute shell commands to validate test coverage.
Evidence: SKILL.md contains instructions to run uv run pytest, uv run nox, and mise run lint. While these are standard development tasks, they represent a direct interface for executing code based on the agent's interpretation of the repository content.

write-python-tests