email-marketing-bible
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill directs the agent to read and process data from external Email Service Provider (ESP) accounts, including subscriber lists, campaign history, and automation flows. This external content serves as a potential vector for indirect prompt injection attacks.
- Ingestion points: Data fetched from ESP APIs (Section 2), subscriber-provided metadata (Section 9), and email bodies from account audits (Section 1).
- Boundary markers: Section 0 defines mandatory "Hard gates," requiring explicit human approval for any send or scheduling action, which mitigates the risk of automated execution of injected instructions.
- Capability inventory: Capability to perform mutating API calls (e.g.,
/send,/trigger), execute shell commands for design logic, and modify flow configurations. - Sanitization: The manual provides an "Anti-slop" linting protocol (Section 4) and a pre-send checklist (Section 3) to verify content before dispatch.
- [COMMAND_EXECUTION]: Section 5 (AI Email Design Protocol) instructs the agent to execute a shell command (
openssl rand -base64 48) to generate random strings used for deriving visual design parameters. - [DATA_EXFILTRATION]: The skill facilitates the transmission of campaign data, contact segments, and marketing content to external platforms, including the author-affiliated Nitrosend and third-party services like Klaviyo, HubSpot, and Mailchimp. These operations are essential to the skill's primary function but involve the handling of sensitive marketing data.
Audit Metadata