requesting-code-review
Pass
Audited by Gen Agent Trust Hub on May 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses
git rev-parseandgit diffto analyze repository history. These are standard developer tools and are used here within their intended scope for auditing code changes. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it interpolates external data (code implementations and plan requirements) into the reviewer subagent's prompt. This could allow malicious content in code or requirements to influence the agent's behavior during the review process.
- Ingestion points: Untrusted data enters the agent context through the
{WHAT_WAS_IMPLEMENTED},{PLAN_OR_REQUIREMENTS}, and{DESCRIPTION}placeholders incode-reviewer.md. - Boundary markers: The template utilizes Markdown headers to demarcate sections but lacks robust delimiters or explicit instructions to the agent to ignore potentially malicious embedded commands.
- Capability inventory: The skill and its subagent have the capability to execute shell commands (
git). - Sanitization: No input sanitization or validation is applied to the content provided in the placeholders before it is processed by the AI.
Audit Metadata