cloudflare-bypass

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). This prompt explicitly shows and instructs embedding proxy credentials and session cookies into command/config templates (e.g., HTTP_PROXY=http://USER:PASS@..., JSON proxy username/password, cf_clearance session use), which means an LLM would be expected to emit secret values verbatim when filling those placeholders.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The content is a high-risk evasion guide that explicitly instructs how to bypass Cloudflare protections (IP/ASN/TLS/fingerprint/JS/CAPTCHA evasion), including use of residential proxies, CAPTCHA solvers, sticky sessions and IP rotation—clearly facilitating deliberate security-evasion and unauthorized scraping.

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

• Third-party content exposure detected (high risk: 1.00). The SKILL.md explicitly instructs the agent to use the browser tool and proxies to navigate and scrape arbitrary Cloudflare-protected public websites (see "Step 2: Use the Browser Tool", the Python Quick Start example, and testing sections), so the agent will fetch and interpret untrusted third‑party web content that could carry malicious instructions.