wiki-ingest

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

• Third-party content exposure detected (high risk: 0.90). The skill's "URL Ingestion" workflow explicitly fetches arbitrary HTTPS pages ("Fetch the page using WebFetch"), saves and parses that web content into .raw/articles/... and then reads/interprets it to create/update wiki pages and drive follow-on actions, which exposes the agent to untrusted, user‑generated or public web content that could carry indirect prompt injections.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill's URL ingestion step explicitly fetches arbitrary user-provided https:// URLs at runtime using WebFetch and then injects the fetched page content into the agent's ingest pipeline (i.e., into model context/prompts), so fetching https://... is a runtime external dependency that can directly control agent behavior.