postplan

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

• Third-party content exposure detected (medium risk: 0.30). The required runtime workflow (“Read” section) takes a user-provided postplan.dev draft URL and uses curl to fetch the /raw contents, so outsider-authored free text in the uploaded HTML is ingested by the LLM at runtime.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs the runtime to fetch arbitrary uploaded HTML from postplan.dev (strip slash, append /raw, curl), and that remote HTML would be used as the document driving agent output, so postplan.dev is a runtime external dependency that can control prompts.