setup

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.80). The list contains a typosquatted domain (bopen.ai.evil.example) that looks crafted to mimic the legitimate bopen.ai site — a common vector for delivering malicious downloads or phishing — while the remaining URLs are local/dev or official/known hosts.

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

• Third-party content exposure detected (medium risk: 0.30). Required runtime workflow POSTs user-controlled selections from the playground UI into /api/plan, where emitPlan(state, body) generates a full instruction prompt; this prompt then gets pasted into an LLM/agent for execution and therefore reflects outsider-authored free text if any of the selection fields or derived state include it.