scraping-recipes

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The prompt explicitly shows proxy credentials embedded verbatim in browser proxy JSON and in HTTP(S)_PROXY environment variables (USER:PASS in URLs), which instructs including secrets directly in commands/configs and creates a high exfiltration risk.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This skill provides explicit, actionable instructions to evade bot detection and paywalls (rotating/sticky residential proxies, cookie clearing, login automation and account rotation) to harvest large volumes of protected or personal data, facilitating unauthorized data exfiltration and account-based abuse.

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

• Third-party content exposure detected (high risk: 0.95). The skill is a web-scraping workflow that uses a browser tool to fetch and render third-party webpages (e.g., Amazon/Google/LinkedIn/Instagram/Zillow/Indeed/News sites) and then extracts readable DOM text into the agent’s LLM context, which is outsider-authored free text from public sites at runtime.