verifying

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

• Third-party content exposure detected (medium risk: 0.65). This skill’s runtime workflow can fetch and read outsider-authored free text from user-supplied links/attachments (e.g., “Links: fetch web body” and “Optional attachments … take priority”), which may contain arbitrary third-party prose that the agent ingests into the LLM context.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill's "Fetch web body" / "Download-the-original" rules explicitly require downloading external documents at runtime (e.g. https://www.imf.org/publications) and loading their contents into the agent context to drive verification instructions, so this is a runtime external dependency that directly controls the agent's prompts.