piu-mcp

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

• Third-party content exposure detected (high risk: 0.80). The skill's execute/verify workflow (bun skills/scripts/piu.ts execute <request_id> and verify <project_id>) resolves env.host + collection.path_prefix + request.url and issues HTTP requests to arbitrary hosts configured in environments, returning status/headers/body that the agent reads and that can influence verification outcomes and subsequent actions—exposing it to untrusted third-party responses.