honcho-integration

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

• Potentially malicious external URL detected (high risk: 0.90). The skill clearly uses the honcho-ai SDK at runtime (e.g., honcho.peer(...).chat(...)) to fetch external content from Honcho's service (see https://app.honcho.dev and https://docs.honcho.dev), and those responses are directly injected into agent prompts, so the Honcho API endpoints are a runtime external dependency that can control prompts.