agent-mail

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

• Third-party content exposure detected (high risk: 0.85). At runtime, the required workflow uses ClawLink/AgentMail actions like agent_mail_get_message and agent_mail_list_messages, which ingest complete email bodies (free-form text) authored by external senders into the agent’s LLM context.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill invokes the npm-hosted CLI at runtime via npx (which fetches and executes remote code) — e.g. https://www.npmjs.com/package/@useclawlink/cli — so it relies on and runs external code during execution.