codex-review-loop

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). The script ingests outsider-authored free text from Codex’s GitHub PR review/comment bodies via gh api .../pulls/$pr/comments (inline findings) and gh api .../issues/$pr/comments (clean verdict body), which are then parsed and included in the LLM context when the agent uses the script output.