sending-transactional-emails
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill implements strong security hygiene by instructing the user to store secret API keys in environment variables and strictly forbidding their inclusion in chat sessions or client-side code.
- [SAFE]: Administrative tools for making email templates transactional utilize a multi-step confirmation flow (prepare/execute with a hash), reducing the risk of accidental or malicious configuration changes.
- [SAFE]: The skill references official PostHog libraries (posthog-node, posthog) and vendor-owned domains (posthog.com), which are consistent with the author's identity and established infrastructure.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data when instructed to extract branding from external websites or email samples. This is a functional requirement managed by the agent's browsing capabilities.
- Ingestion points: User-provided URLs and email HTML/screenshots (SKILL.md, Step 2).
- Boundary markers: Not explicitly defined for the branding extraction phase.
- Capability inventory: Tools for email template design, sender verification, and transactional sending (SKILL.md, Steps 3, 5, 8).
- Sanitization: Relies on agent-side parsing of extracted brand attributes.
Audit Metadata