uv-strict-python
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes a 'scripts/check.py' template that manages tool execution via 'subprocess.run'. The implementation is secure, using static argument lists and avoiding shell interpolation, which prevents command injection.
- [INDIRECT_PROMPT_INJECTION]: The skill proactively mitigates indirect prompt injection by providing normative agent guidelines. Ingestion points include untrusted repository files and tool outputs (referenced in 'coding-standard.md'); boundary markers are implemented via delimiter-bounded blocks; capabilities are restricted to hardcoded tool suites; and sanitization is enforced through strict agent instructions to treat data as non-authoritative.
- [EXTERNAL_DOWNLOADS]: The 'templates/check.yml' workflow references the 'astral-sh/setup-uv' GitHub Action. It follows supply chain security best practices by pinning the dependency to a specific commit SHA, ensuring build integrity.
- [DYNAMIC_EXECUTION]: Bash shims for 'python' and 'pip' are provided to intercept direct commands and redirect agents toward managed 'uv' execution, preventing accidental use of insecure or unconfigured system environments.
Audit Metadata