crawl4ai-web-crawler

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill examples show API tokens and credentials embedded as string literals (e.g., api_token="your-key"/"your-api-key" and proxy="http://user:pass@..."), which encourages passing secrets verbatim in code/requests and would require an LLM/agent to handle or output secret values directly, creating an exfiltration risk.

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

• Third-party content exposure detected (medium risk: 0.30). SKILL.md and references/api-reference.md describe runtime behavior where the user supplies a target url (e.g., crawler.arun(url="...")), and the crawler then ingests the fetched page’s free-form text/HTML to generate result.markdown and optionally feed it into LLMExtractionStrategy/LLMExtractionStrategy.instruction—so outsider-authored page content can be injected.