alibaba-image-models
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides code and instructions for downloading generated image assets from Alibaba Cloud endpoints. The implementation incorporates several security best practices, including enforcing HTTPS, validating that hostnames resolve to public IP addresses to prevent Server-Side Request Forgery (SSRF), and performing strict integrity checks on downloaded files (e.g., verifying PNG magic bytes and limiting pixel dimensions to prevent resource exhaustion).
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-provided text prompts and image references. It mitigates injection risks by instructing the agent to use gated iteration loops, requiring human review for deceptive or sensitive context, and providing clear guidance on handling moderation failures (DataInspectionFailed) without attempting to obfuscate requests.
- [SAFE]: Secret management is handled via environment variables (e.g., DASHSCOPE_API_KEY) and the skill explicitly warns against placing credentials in source code, prompts, or logs. The documentation and provided logic reflect a strong security-first approach to productionizing AI workflows.
Audit Metadata