peer-review-loop

Pass

Audited by Gen Agent Trust Hub on Apr 9, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local shell scripts scripts/codex-review.sh and scripts/setup-build.sh to manage the Ralph Loop lifecycle and review iterations.
  • [COMMAND_EXECUTION]: It invokes the codex command-line utility for performing automated peer reviews.
  • [EXTERNAL_DOWNLOADS]: The skill relies on the @openai/codex package from the public NPM registry, which is a well-known service from a trusted provider.
  • [PROMPT_INJECTION]: The skill possesses an Indirect Prompt Injection surface through the processing of external data. * Ingestion points: Reads project source code and Cavekit definition files from the context/kits/ directory. * Boundary markers: There is an absence of explicit delimiters or instructions to ignore embedded commands when these files are interpolated into the builder and reviewer prompts. * Capability inventory: The skill has the capability to execute shell scripts and modify the persistent .mcp.json configuration file based on the loop's output. * Sanitization: No sanitization or validation logic is described for the data ingested from files or the findings generated by the peer reviewer before they are processed.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 9, 2026, 07:13 PM