newsjack-triage

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

• Third-party content exposure detected (high risk: 0.85). The skill ingests targeted_candidates.json from origin-apply, whose evidence[].excerpt is runtime-provided free text from external news sources (e.g., editorial/newswire articles or other public coverage) and is read into the LLM context for provenance/standing decisions.