agentmail

Pass

Audited by Gen Agent Trust Hub on Mar 6, 2026

Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection vulnerability surface. * Ingestion points: Incoming email content is retrieved via the AgentMail API as shown in the SKILL.md code snippets. * Boundary markers: There are no delimiters or 'ignore embedded instructions' markers present in the processing examples. * Capability inventory: The agent has the ability to send emails (mail.messages.send) and register webhooks. * Sanitization: No input validation or content filtering of incoming messages is implemented in the logic.
  • [DATA_EXFILTRATION]: External API Communication. The skill communicates with the vendor's infrastructure at api.theagentmail.net to manage accounts and process email data.
  • [EXTERNAL_DOWNLOADS]: Vendor SDK Dependency. The skill references the vendor-owned Node.js package @agentmail/sdk for its functionality.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 6, 2026, 01:40 AM