discover

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly ingests external content in the document_phase (accepting URLs) and research_phase (querying third-party MCP servers like perplexity, context7, and firecrawl), which the agent is instructed to read and use to influence elicitation, synthesis, and follow-up actions, exposing it to untrusted third-party content that could carry indirect prompt injection.