review-work
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The code-reviewer subagent template processes potentially untrusted external data via placeholders. * Ingestion points: Data is injected into {WHAT_WAS_IMPLEMENTED}, {PLAN_OR_REQUIREMENTS}, and {DESCRIPTION} within references/code-reviewer.md. * Boundary markers: The template lacks specific delimiters or instruction-guarding warnings to separate external data from the agent's core instructions. * Capability inventory: The subagent is authorized to run git commands and provide a definitive merge recommendation. * Sanitization: The skill does not appear to sanitize or validate the external content before interpolation.
- [COMMAND_EXECUTION]: The skill utilizes shell commands to manage and analyze the local git repository. * Evidence: SKILL.md and references/code-reviewer.md instruct the agent to use git merge-base, git rev-parse, and git diff.
Audit Metadata