1688

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This skill is explicitly designed to automate large-scale harvesting of personal and corporate contact information (phones, emails, legal representative names) from 1688 and Google while instructing evasion techniques (avoiding reCAPTCHA, batching/delays, bypassing hidden phone protections), which constitutes deliberate data exfiltration and a high risk for abuse (spam, unsolicited contact, lead harvesting, doxxing).

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

• Third-party content exposure detected (high risk: 0.85). Outsider free text is ingested when the skill navigates to Google Search results and then runs evaluate on the loaded page DOM (public web content authored by third parties), extracting h3 titles and a[href] URLs into the agent context.