web-scraping-proxy

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The prompt repeatedly shows and constructs proxy credentials inline (e.g., USER:PASS in JSON, HTTP_PROXY env var, and f"http://{proxy_user}:{proxy_pass}@..."), which encourages the agent to ask for and embed secrets verbatim into commands/connection strings, creating an exfiltration risk.

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly instructs the agent to fetch and parse open/public websites (e.g., "Step 1: Check Protection Level" with curl to https://target-site.com and "Step 5: Rotate and Scrape" using requests.get) and lists scraping social media, news sites, and other public pages whose untrusted, user-generated content is read and used to choose tools and next actions (browser vs HTTP, delays, sticky sessions), creating a clear path for indirect prompt injection.