anti-ui-slop

Pass

Audited by Gen Agent Trust Hub on Aug 24, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the agent to browse and search a public catalogue of UI design screens at https://uizze.com to collect design evidence. This is used for reference purposes and does not involve the automated execution of remote scripts.- [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface as it ingests information from both the external design catalogue and the local repository codebase to inform its output. However, the workflow includes structured checkpoints to maintain safety.
  • Ingestion points: Data enters the agent context through searching the external uizze.com catalogue and reading files within the local repository.
  • Boundary markers: The instructions mandate the creation of a 'Design Contract' to explicitly document decisions and a 'Finish Gate' checklist to verify the resulting UI code against specific product requirements.
  • Capability inventory: The skill allows the agent to read the repository and write/refactor UI code, but it does not specify tool usage for shell execution or network writes.
  • Sanitization: The handoff format requires a verification section and a risk assessment, ensuring the agent evaluates the generated content before completion.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 24, 2026, 05:49 PM
Security Audit — agent-trust-hub — anti-ui-slop