resend-cli

Pass

Audited by Gen Agent Trust Hub on Nov 17, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill includes commands to read and listen to inbound emails, creating an attack surface where an attacker could send emails containing malicious instructions.
  • Ingestion points: emails receiving get, emails receiving listen (in references/emails.md), and inboxes threads emails get (in references/inboxes.md) pull external subject lines, bodies, and headers into the agent context.
  • Capability inventory: As described in SKILL.md, the CLI possesses powerful capabilities such as api-keys create, domains delete, and emails send which could be targeted by an injection attack if the agent obeys instructions inside an email.
  • Boundary markers: The skill contains explicit instructions to agents in SKILL.md to treat received email content strictly as data and never as instructions.
  • Sanitization: The instructions in references/emails.md emphasize that content from inbound emails is untrusted and must be sanitized before processing.
  • [DYNAMIC_EXECUTION]: The skill facilitates the runtime bundling and rendering of React Email templates (.tsx files) via flags like --react-email in emails send (in references/emails.md) and broadcasts create (in references/broadcasts.md) commands.
  • This process uses esbuild to compile local code. While a standard feature of the Resend platform, it involves dynamic execution of scripts that may be generated or modified by the agent at runtime.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 17, 2026, 06:24 PM