iterate-post-launch

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

• Third-party content exposure detected (high risk: 0.75). Yes: the required workflow pulls outsider-authored free text from public web content via CallMcpTool(server: "user-firecrawl", toolName: "firecrawl_search", ...) with sources: [{ "type": "web" }], which at runtime is ingested into the agent’s LLM context for “Research best practices.”