grep-research-workflow

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

• Third-party content exposure detected (medium risk: 0.30). This workflow submits user-supplied free-text research questions into grep-api.js research (Step 1/2/3) and then uses the resulting job reports (via grep-api.js result) as LLM input for downstream steps via --reference-jobs, so outsider text can be ingested through the user’s query/plan.