knowledge-absorber

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

• Third-party content exposure detected (high risk: 0.90). The skill's scripts/content_ingester.py explicitly fetches and renders arbitrary public URLs (using requests and DrissionPage) into .trae/skills/knowledge-absorber/config/raw_content.txt, and SKILL.md / references/system_prompt.md require the agent to read and act on that raw_content.txt to generate behavior, so untrusted third-party page content can materially influence the agent.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's scripts/content_ingester.py fetches an arbitrary user-supplied URL at runtime (requests.get(args.input)) and writes the fetched content to .trae/.../raw_content.txt which is then injected into the agent's context for generation, so a remote URL provided at runtime can directly control the model's input/output (flagging the user-supplied URL fetched by scripts/content_ingester.py).