code-ultrareview

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

• Third-party content exposure detected (high risk: 0.85). Intent-axis --reconcile can ingest outsider-authored free text from GitHub PR/issue bodies via scripts/derivation/run.py → fetch_pr_body_text() / fetch_issue_body_text() (uses gh api/gh pr view), and that text is then passed into the Intent subagent’s LLM context for classification.

MEDIUM W021: Hidden or invisible Unicode characters detected (potential obfuscation or prompt injection).

• Hidden Unicode characters detected (1 type(s) found)