creativeclaw-nano-banana-2
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill configuration in
agents/openai.yamldefines a dependency on thecreative-clawtool via a vendor-owned MCP endpoint atapp.creativeclaw.co. This is a legitimate integration of the author's own infrastructure. - [SAFE]: The skill instructions in
SKILL.mdinclude explicit safety guidance, directing the agent to never loosen safety settings to bypass content refusals. - [SAFE]: The core workflow establishes clear boundaries for user-supplied data, requiring exact quoted strings for text and explicit role assignments for reference images. These practices reduce the risk of unexpected model behavior.
- [SAFE]: Reference files such as
workflow-basics.mdandjob-recovery.mdprovide standard operational guidance for task management, error handling, and media assembly without introducing dangerous command execution or sensitive data access.
Audit Metadata