subject-line-lab
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted subject lines and preheaders from external Email Service Providers (ESPs) during the historical performance mining phase.
- Ingestion points: Step 1 pulls data from Klaviyo, Mailchimp, Rule, or Get a Newsletter.
- Boundary markers: The instructions lack delimiters or constraints to prevent the model from executing any commands found within the processed subject lines.
- Capability inventory: The skill has access to file writing and vendor-specific tools for persisting findings and context nodes.
- Sanitization: No input validation or sanitization is performed on the retrieved subject lines before they are processed by the agent.
Audit Metadata