The Agent Skills Directory

[SAFE]: The skill provides clear and structured instructions for using Mapbox Model Context Protocol (MCP) tools, such as ground_location_tool and directions_tool, to ensure responses are based on live data rather than potentially stale training data.
[SAFE]: Analysis of all instructions and eval cases shows no evidence of prompt injection, data exfiltration, or credential harvesting. The skill correctly instructs the agent to handle user coordinates for the sole purpose of location search and routing.
[SAFE]: No obfuscation, hidden URLs, or unauthorized remote code execution patterns were found. All referenced services are well-known Mapbox APIs consistent with the skill's stated purpose.
[SAFE]: The skill implements security and reliability best practices by requiring explicit citations of live data sources and discouraging the fabrication of business details.

mapbox-location-grounding