keyword-research

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly calls the Sorftime MCP APIs (e.g., SKILL.md Step 1.1 / references/api-tools-reference.md and scripts/keyword_collector.py) to fetch product keywords, product_detail and product_reviews (third‑party/user-generated data) and then feeds those results into an LLM classification prompt / automatic LLM classification (--claude-code-env and the classification prompt template), so untrusted external content is ingested and directly used to drive decisions like negative-keyword lists and campaign actions.