gh-review-suggestions

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

• Third-party content exposure detected (medium risk: 0.65). The workflow runs gh pr diff <pr> > /tmp/pr.diff and then reads that PR-authored diff text via diff-line-position.mjs (readFileSync(path, 'utf8')), so outsider free-form PR content is ingested into the agent’s LLM context indirectly through the tool outputs/used review text.