wecomcli-email

Pass

Audited by Gen Agent Trust Hub on Sep 10, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external and potentially untrusted email content, which serves as a vector for indirect prompt injection. \n
  • Ingestion points: Email bodies and attachments are retrieved via the wecom-cli mail get command (referenced in references/get-mail.md). \n
  • Boundary markers: The skill defines explicit defensive rules in references/security.md to ignore any directive text found in email bodies and to add a 'Request Source Reminder' if recipient addresses are extracted from email content. \n
  • Capability inventory: The skill possesses the capability to send emails to arbitrary addresses (wecom-cli mail send) and download media files (wecomcli-media). \n
  • Sanitization: The instructions in references/security.md and references/send-mail.md mandate that the agent refuse to include <script> tags, event handlers, or malicious URI schemes in outgoing content. \n
  • [COMMAND_EXECUTION]: The skill performs its operations by executing the wecom-cli binary. This tool is a standard part of the vendor's ecosystem used to interact with Enterprise WeChat APIs.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 10, 2026, 07:05 AM
Security Audit — agent-trust-hub — wecomcli-email