requesting-code-review

Pass

Audited by Gen Agent Trust Hub on Jun 18, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The code-reviewer.md template exhibits a vulnerability to indirect prompt injection by interpolating potentially untrusted data directly into executable command blocks.
  • Ingestion points: The template processes data through placeholders including {BASE_SHA}, {HEAD_SHA}, and {DESCRIPTION}.
  • Boundary markers: The template lacks delimiters or explicit instructions to the agent to treat interpolated data as non-executable content within the shell command block.
  • Capability inventory: The subagent is instructed to use the git diff command, which involves subprocess execution.
  • Sanitization: No validation or escaping mechanisms are implemented to ensure that the interpolated strings do not contain shell metacharacters or malicious commands.
  • [COMMAND_EXECUTION]: The reviewer subagent is instructed to execute shell commands (git diff) using parameters derived from variables that may originate from external sources like pull request metadata, creating a path for arbitrary command execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 18, 2026, 11:32 PM
Security Audit — agent-trust-hub — requesting-code-review