mapbox-style-quality

Pass

Audited by Gen Agent Trust Hub on Apr 1, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns or security risks were identified. The skill is educational and focuses on style validation and accessibility standards.
  • [DATA_EXFILTRATION]: No evidence of hardcoded credentials or unauthorized data transmission. External links target official documentation and well-known standards.
  • [PROMPT_INJECTION]: The skill processes external style and GeoJSON data, creating an ingestion surface for indirect prompt injection. 1. Ingestion points: Mapbox Style JSON and GeoJSON sources mentioned in SKILL.md. 2. Boundary markers: Absent. 3. Capability inventory: Mapbox-specific validation, comparison, and optimization tools. 4. Sanitization: The skill mandates the use of validate_geojson_tool and validate_expression_tool to sanitize and verify all incoming data.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 1, 2026, 08:10 AM