subagent-driven-review

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

• Third-party content exposure detected (high risk: 0.85). The required workflow uses RC native tools (rp_search, rp_abstracts, rp_fulltext) that fetch OpenAlex/Elsevier OA abstracts/full text at runtime and then pass those retrieved prose records into subagent prompts for scoring/extraction, which is outsider-authored content (publisher/author text) flowing into the LLM context.