amazon-review-intelligence-extractor

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

• Third-party content exposure detected (medium risk: 0.30). In SKILL.md’s “Insufficient Data Fallback” workflow, the agent fetches raw Amazon review bodies via zoodata.py reviews-raw and then uses the caller LLM to tag and reduce those outsider-authored texts (review-tag-prompt / review-reduce-prompt), so free text from review authors is directly ingested by the LLM at runtime.