code-review-ai-ai-review

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

• Third-party content exposure detected (high risk: 0.85). The workflow ingests PR-provided diff/description (e.g., Review: **$ARGUMENTS** and **Modified Code:** {code_diff} / {pr_description}) into the LLM prompt; those texts are authored by outsiders (other contributors on the pull request) and become readable free text in the runtime prompt.