mena-marketing-localization
Pass
Audited by Gen Agent Trust Hub on Jun 30, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill instructions explicitly prohibit high-risk automated actions such as launching ads, publishing content, or uploading audience lists without human approval. It enforces privacy by directing the agent to protect credentials, pixels, and customer data.
- [SAFE]: All external references and downloads in
sources.ymltarget official government domains (e.g., .gov.sa, .gov.eg, .ma, u.ae) or well-known platform documentation (Google, Meta, TikTok, LinkedIn, GSMA). These are verified sources for marketing and regulatory information. - [SAFE]: The skill implements a structured confidence and evidence model, requiring the agent to distinguish between source-backed facts, hypotheses, and unknowns, which reduces the risk of hallucination or misinformation in localized content.
- [SAFE]: Multi-layered review gates (native, legal, platform, and operational) are mandated for all sensitive or public-facing content, ensuring that cultural nuances and legal requirements are handled by experts rather than autonomously by the AI.
Audit Metadata