mapbox-mcp-devkit-patterns

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions for the user to download and configure the Mapbox MCP DevKit server from Mapbox's official GitHub repository and to connect to their hosted service at mcp-devkit.mapbox.com using npx.
  • [INDIRECT_PROMPT_INJECTION]: The skill directs the agent to ingest and process external, potentially untrusted data such as GeoJSON, Mapbox style JSON, and paint/layout expressions through its toolset.
  • Ingestion points: Untrusted data enters the agent context through the validation and creation tools (e.g., validate_style_tool, validate_geojson_tool) as described in SKILL.md and references/workflows.md.
  • Boundary markers: The skill does not provide specific instructions to the agent to use boundary markers or to ignore potential instructions embedded within the data payloads.
  • Capability inventory: The skill enables the agent to perform actions on a user's Mapbox account, including creating/modifying styles and generating access tokens.
  • Sanitization: No sanitization or filtering logic is prescribed in the skill's instructions; the security relies on the underlying MCP server's implementation and the model's internal guardrails.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 12:03 AM
Security Audit — agent-trust-hub — mapbox-mcp-devkit-patterns