restore-and-upscale

Pass

Audited by Gen Agent Trust Hub on Jul 2, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns, obfuscation, or unauthorized access attempts were detected in the skill's instructions or examples. The functionality is consistent with its stated purpose.
  • [DATA_EXFILTRATION]: The skill involves sending user-provided image and video URLs to the Runware processing infrastructure (runware.ai). This is the intended operational behavior for the image processing services provided by the vendor.
  • [PROMPT_INJECTION]: The skill processes external media content via URLs, which represents a potential surface for indirect prompt injection. However, it incorporates security best practices by instructing the agent to validate model inputs and configurations against live schemas before execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 2, 2026, 04:08 PM
Security Audit — agent-trust-hub — restore-and-upscale