knowledge-ingest-source

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

• Third-party content exposure detected (high risk: 0.85). The skill’s required workflow includes a runtime webfetch path for --url that fetches arbitrary public web pages/PDFs and passes their extracted body text into the agent via the shared cache that source-ingester later reads with fetch-cache.py fetch (indirect prompt injection risk from outsider-written free text).