docs-auditor

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The skill explicitly mandates stealthy web searches ("Do this search silently") and auto-fetching of many derived URLs treating them as "user-provided" (without asking), which enables hidden network requests and reconnaissance that could be abused for data collection or exfiltration even though no explicit payload/execution code is present.

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

• Third-party content exposure detected (high risk: 0.95). The skill fetches outsider-authored free text from the target docs site at runtime—e.g., it derives and fetches {base}/robots.txt, {base}/sitemap.xml, {base}/llms.txt, and multiple HTML pages, then ingests their contents into the LLM context for evaluation (public web content controlled by the docs publisher, not the operating user).