ai-interface-reviewer

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches AI principles, summaries, and business impact data from the author's official domain.
  • Evidence: API calls are directed to https://uxuiprinciples.com/api/v1/principles.
  • [COMMAND_EXECUTION]: The skill defines tools that utilize the curl command to interact with its backend service.
  • Evidence: toolbox.lookup_ai_principle and toolbox.list_ai_principles both use curl with user-provided or environment-stored API keys.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted descriptions of external interfaces to generate audit reports.
  • Ingestion points: Interface descriptions provided by the user in the agent context.
  • Boundary markers: The skill instructs the agent to classify and evaluate based on a specific JSON schema, but does not specify delimiters for the input data.
  • Capability inventory: Network access via curl to uxuiprinciples.com.
  • Sanitization: The skill relies on LLM reasoning to map descriptions to principle slugs for API lookups.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 04:21 PM
Security Audit — agent-trust-hub — ai-interface-reviewer