concurrent-cached-fetch

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

• Third-party content exposure detected (high risk: 0.75). The skill’s runtime workflow is to perform public external HTTP fetches (outsider-authored web content) and then ingest the fetched response body as readable text/JSON into the agent’s context via the tool’s returned results (e.g., cached GET responses are parsed and returned for LLM use), creating an indirect prompt-injection path from arbitrary web pages into the LLM context.