03-performance-eval-global
Pass
Audited by Gen Agent Trust Hub on May 19, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No prompt injection or behavior override patterns were detected. The instructions follow standard marketing diagnosis procedures.
- [SAFE]: Data exposure and exfiltration checks are negative. The skill references official Model Context Protocol (MCP) integrations for Meta, Google, and TikTok, which are well-known services.
- [SAFE]: No obfuscation techniques, hidden URLs, or encoded commands were found in the skill or its variants.
- [SAFE]: The skill does not download or execute remote code. It utilizes established platform integrations to pull performance metrics.
- [SAFE]: Privilege escalation and persistence mechanisms are absent. The skill operates within the expected scope of a data analysis agent.
- [SAFE]: The diagnostic framework and regional variants (US, EU, SEA, LATAM) provide legitimate marketing context and benchmarks consistent with the skill's stated purpose.
Audit Metadata