producing-threat-intelligence

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s runtime workflow “Producing Threat Intelligence” includes reading external sources by fetching arbitrary public webpages as Markdown (via curl -sL "https://defuddle.md/<url>"), which means outsider-authored free text from those URLs is ingested by the LLM after an attacker-supplied URL is supplied in the fetch path.