requesting-code-review
Pass
Audited by Gen Agent Trust Hub on Jun 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell commands (git) for version control operations. The prompt template in
code-reviewer.mdinterpolates{BASE_SHA}and{HEAD_SHA}directly into shell commands. This creates a surface for command injection if an agent populates these placeholders with unvalidated or malicious strings.- [PROMPT_INJECTION]: The code review agent ingests output fromgit diff, which is a source of untrusted data that could contain malicious instructions designed to influence the agent's behavior. - Ingestion points: External code content is read into the agent context via
git diffas defined incode-reviewer.md. - Boundary markers: The template lacks clear instructions or markers to distinguish between the diff data and the instructions, increasing susceptibility to indirect prompt injection.
- Capability inventory: The agent has the capability to execute shell commands and generate technical reports.
- Sanitization: There is no evident sanitization of the code diff output before it is processed by the agent.
Audit Metadata