generating-images
Pass
Audited by Gen Agent Trust Hub on Aug 22, 2026
Risk Level: SAFE
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill interfaces with image generation services via standard platform tools (image_generate). This behavior is consistent with its stated purpose of routing prompts to AI models and does not involve unauthorized code execution.
- [DATA_EXPOSURE]: The skill manages user-supplied text and images for the purpose of generating visual content. It does not attempt to access sensitive system files, environment variables, or credentials. It correctly guides users to configure their own API keys via the platform if provider errors occur, adhering to secure credential management practices.
- [PROMPT_INJECTION]: The instructions explicitly direct the agent to uphold safety guidelines when processing user requests. If a request violates safety policies (e.g., NSFW or exploitative content), the skill instructs the agent to inform the user and suggest acceptable alternatives rather than attempting to bypass filters or override system constraints.
- [COMMAND_EXECUTION]: No evidence of arbitrary shell command execution, persistence mechanisms, or unauthorized file system operations was found. The skill operates entirely through natural language instructions and predefined tool calls.
Audit Metadata