solid-principles
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted code and test files from the repository to perform reviews and refactoring, creating a surface where malicious repository content could influence agent behavior.\n
- Ingestion points: Source code, tests, usage sites, and dependency registrations are read from the repository (SKILL.md, references/review-checklist.md).\n
- Boundary markers: Instructions require the agent to confirm authorization for edits and to explicitly report which tests were run vs not run (SKILL.md, references/verification.md).\n
- Capability inventory: The agent is empowered to modify files (SKILL.md) and execute repository-defined test/build commands (SKILL.md, references/verification.md).\n
- Sanitization: No specific input sanitization or filtering for ingested code content is defined.\n- [COMMAND_EXECUTION]: The skill instructs the agent to identify and execute build and test commands from the local repository environment to verify design changes.\n
- Evidence: 'Identify test/build commands from the repository; never fabricate results' (SKILL.md).\n
- Evidence: 'Exercise focused tests plus relevant broader checks if available. Explicitly report commands, pass/fail/not run...' (SKILL.md).
Audit Metadata