quality-assessment

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

• Third-party content exposure detected (high risk: 0.70). The skill (via subagent-spawning/spawn-agent) passes paper_details generated by upstream “paper-research or paper-search”, whose contents are commonly derived from outsider-authored public web/recorded paper text, creating a runtime path where that free-form paper text can enter the subagent prompt context (indirect prompt injection risk).