claude-cli

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

• Third-party content exposure detected (high risk: 0.75). SKILL.md instructs the runtime to fetch and execute outsider-authored free text from public URLs (e.g., curl -fsSL https://claude.ai/install.sh | bash and similar install scripts), which can be ingested as readable script text during execution.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill instructs running remote install scripts that are fetched and executed at runtime (e.g., curl -fsSL https://claude.ai/install.sh | bash, irm https://claude.ai/install.ps1 | iex, and curl -fsSL https://claude.ai/install.cmd ... && install.cmd), which directly executes remote code and is presented as the recommended/native install path.