gpt-image-2-prompting
Pass
Audited by Gen Agent Trust Hub on May 31, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is strictly limited to text-to-text prompt transformation and structural optimization. It does not perform any network operations, external downloads, or file system modifications beyond its own directory.
- [SAFE]: The logic regarding the
image_generatetool is implemented with a clear authorization flow, requiring the agent to ask for user permission before calling the tool, which prevents unauthorized usage or cost escalation. - [SAFE]: Analysis of the extensive reference files (templates, examples, and themed prompt lists) confirms they contain only benign creative content and instructional structures intended to guide the LLM's output format.
- [SAFE]: There is no evidence of obfuscation, credential harvesting, or attempts to bypass agent safety guidelines. The instructions reinforce structural adherence and clarity in communication.
Audit Metadata