cold-email-sequence
Pass
Audited by Gen Agent Trust Hub on Jul 29, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues detected.
- [DATA_EXPOSURE]: The skill reads from a local file
.agents/product-marketing.md. This is standard behavior for an agent to retrieve project-specific context and does not involve exfiltration or unauthorized access. - [EXTERNAL_DOWNLOADS]: Includes a reference to a marketing consultancy website (aajconsult.com) for additional tools, which is provided as a static link for informational purposes.
- [PROMPT_INJECTION]: The skill ingests data from local project files. While this represents a surface for indirect prompt injection, the skill includes guardrails requiring the agent to verify claims against supported proof and focuses purely on text generation rather than command execution.
Audit Metadata