cn-image-gen
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides documentation and code snippets for downloading and loading pre-trained model weights from official repositories on Hugging Face (e.g., THUDM, Kwai-Kolors, tencent). These are well-known organizations and the operations are standard practices for AI model deployment.
- [CREDENTIALS_UNSAFE]: All code snippets in the skill correctly use placeholders for sensitive information (e.g.,
YOUR_API_KEY,YOUR_KEY,YOUR_ARK_API_KEY,API_KEY) rather than hardcoding actual secrets, adhering to secure coding practices. - [COMMAND_EXECUTION]: The skill contains example
curlcommands and Python scripts intended for API interaction and local inference using thediffuserslibrary. These are provided as educational examples and do not include malicious or unauthorized command execution patterns. - [INDIRECT_PROMPT_INJECTION]: While the skill provides prompt templates for user-directed image generation, it does not include execution capabilities that could be exploited via untrusted data. The skill's primary purpose is informational, and it does not ingest external data into sensitive execution contexts.
Audit Metadata