Geo-sleuth is an agent skill that determines the geographic location where a photograph was taken by analyzing visual evidence and displaying its reasoning process.
The tool solves the problem of photo geolocation by combining multiple data sources and analysis techniques. It examines OpenStreetMap geometry, elevation profiles, satellite imagery, and street view data to pinpoint camera position and the direction the camera was facing. The agent works by reading a skill definition file and executing Python scripts that process the image and cross-reference it against these geographic datasets. The output includes the calculated camera position, viewing direction, and satellite imagery as evidence of the analysis.
Developers should adopt this tool if they are building AI agent applications that need geolocation capabilities and want the agent to show its work transparently. It suits projects where understanding how a location was determined matters as much as the result itself. The tool integrates with multiple agent platforms including Claude Code, Codex, Cursor, Gemini CLI, OpenCode, and GitHub Copilot, making it accessible across different development environments. Anyone using an agent framework that reads skill definition files and executes shell commands can use this tool.
The project maintains an active contribution posture with a contributing guide in place. Documentation is available in multiple languages, indicating attention to accessibility for international users. The codebase requires Python and includes a diagnostic script for users to verify their environment meets all dependencies before first use.
