claude-cli-docs

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.80). Several URLs point directly at installer scripts (https://claude.ai/install.sh, .ps1, .cmd) which are high-risk to fetch-and-execute (curl|bash, irm|iex) without verification; the remaining links are documentation or official pages and appear benign.

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

• Third-party content exposure detected (high risk: 0.85). SKILL.md is a reference router that instructs the agent to load documentation pages (e.g., installation/cli-reference/hooks/skills/workflows) from the web at runtime, so the LLM context can include outsider-authored free text fetched from public URLs.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly instructs the agent to WebFetch the official docs at runtime (e.g., https://code.claude.com/docs/en/... and the full index https://code.claude.com/docs/llms.txt) so remote page content is injected into the model context and can directly influence prompts/responses.