age
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted code diffs, pull request metadata, and project files as its primary input for analysis. 1. Ingestion points: Git diffs, PR bodies, and scoped file paths defined in SKILL.md. 2. Boundary markers: No specific boundary markers or instructions to isolate untrusted content from the system prompt are defined. 3. Capability inventory: Execution of local bundled scripts (age.pyz), dispatch of sub-agents for specialized review tasks, and file writes to the .cheese/ directory. 4. Sanitization: No explicit sanitization of external content before processing is documented.
- [COMMAND_EXECUTION]: The skill relies on a local Python Zip Application (skills/age/scripts/age.pyz) and external CLI tools such as git and gh to manage repository data, resolve artifacts, and execute the review lifecycle.
- [DYNAMIC_EXECUTION]: The skill dispatches parallel sub-agents (reviewer role) to handle different dimensions of the code review, utilizing a fan-out architecture in read-only, fresh-context environments as described in references/fan-out.md.
Audit Metadata