47-design-review

Pass

Audited by Gen Agent Trust Hub on Aug 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill's instructions are focused entirely on providing a quality review framework for designers. There are no signs of base64 obfuscation, base64-encoded commands, or hidden control characters.
  • [DATA_EXFILTRATION]: While the skill prompts for external resources ("Upload file / screenshot / share link"), this is a necessary functional requirement for design evaluation. The skill does not attempt to access sensitive system files like SSH keys or cloud credentials, nor does it exfiltrate data to unauthorized domains.
  • [PROMPT_INJECTION]: The skill processes untrusted external data (images or links) which could serve as a vector for indirect prompt injection. However, the skill has a very limited capability set (generating text feedback) and lacks access to dangerous tools.
  • Ingestion points: Designs provided via upload, screenshot, or share link as described in the 'Thu thap thong tin' section of SKILL.md.
  • Boundary markers: Absent; the instructions do not explicitly tell the agent to ignore text within the designs.
  • Capability inventory: The skill only generates markdown text or in-chat feedback.
  • Sanitization: None specified for input data.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 16, 2026, 07:05 AM
Security Audit — agent-trust-hub — 47-design-review