ai-cmo-os
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The system processes untrusted data from external sources while possessing the authority to execute high-impact tasks, creating a potential vector for indirect prompt injection.\n
- Ingestion points: Data from web research (Google, Bing, SerpApi), social media (Meta, TikTok), and ecommerce (Shopify) platforms.\n
- Boundary markers: The skill describes "Context Boundary Enforcement" and the use of state snapshots to maintain isolation between tasks.\n
- Capability inventory: The system can perform file modifications, API calls for campaign management, and code commits to production branches.\n
- Sanitization: The architecture utilizes "Tier 2 Gated Tasks" for human review of high-risk actions and implements output validation to ensure quality and alignment.\n- [EXTERNAL_DOWNLOADS]: The skill documentation describes integrations with established technology and marketing platforms such as AWS, Google, Meta, Shopify, and TikTok. These references represent standard external service connections for the intended business use case and utilize official, well-known domains.
Audit Metadata