he-spec

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

• Third-party content exposure detected (high risk: 0.75). The required workflow includes “Detect External Input” where the user may provide a URL or pasted external spec/PRD, and the skill then “Read/parse the external document” / “Extract relevant details from the artifacts,” which can ingest outsider-authored free text from public web content or third-party documents into the agent’s LLM context.