minimizing-ty-ecosystem-changes
Pass
Audited by Gen Agent Trust Hub on Jun 23, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches the mypy-primer utility directly from its official GitHub repository via the uv package manager. This is the standard tool for regression testing in the Python typing community and is used here for its intended functionality.- [COMMAND_EXECUTION]: The skill instructs the agent to compile the ty binary using cargo and run various analysis scripts locally. These commands are necessary for the skill's primary function of comparing software versions and do not involve unauthorized privilege escalation.- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it processes external code from third-party Python projects. Evidence: 1. Ingestion: External project code is fetched via scripts/setup_primer_project.py; 2. Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the provided text; 3. Capability inventory: The skill builds and executes the ty binary against the fetched code; 4. Sanitization: No sanitization of the fetched project code is documented. However, the risk is minimal as the code is used as input for a static analysis binary rather than being directly interpreted by the agent's logic.
Audit Metadata