enrich-waterfall

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This skill is high-risk: it is explicitly designed to harvest personal contact details (emails, phone numbers) and scrape social profiles, with paid "escalate" steps that reveal personal contact info — functionality readily abused for mass data harvesting or targeted doxxing.

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

• Third-party content exposure detected (medium risk: 0.30). This skill ingests caller-supplied free text parameters (e.g., name, company, linkedinUrl) and sends them into LLM-relevant enrichment/search endpoints (including Brave news-search query strings) without requiring the agent to first select a specific item.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 1.00). The skill explicitly compiles manifest steps into selat-pay calls and settles them via the SELAT Router / Circle Gateway, includes per-step monetary caps and live prices, and shows selat-pay POST examples with a 402 preflight and settlement chain tied to a funded Circle Gateway balance. These are specific payment gateway / settlement operations (direct financial execution), not generic API or browser automation.