copaw-ai-assistant

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md includes a "Skill with async support" example (skills/summarize_url.py) that uses httpx to fetch and return the content of arbitrary public URLs, so the agent will read/unpack untrusted third‑party web content which could inject instructions affecting its actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The summarize_url skill performs an httpx.AsyncClient.get at runtime on an arbitrary, user-supplied URL (the "url" argument fetched via httpx.get(url)) and returns resp.text into the assistant context, so remote content can directly control prompts.