Skip to content
geo-sleuth

geo-sleuthSkill

by Oldcircle
Released
1 views
MIT
Repository
markdown Development

Summary

Find where a photo was taken: EXIF, OCR, OpenStreetMap, skyline and satellite matching, with coordinates, error radius and evidence images.

Features

  • EXIF, OCR and reverse image search in one intake command
  • Scripted candidate board ranks locations so the model judges only the top few
  • OpenStreetMap geometry, elevation skyline rendering and sun/shadow maths
  • Satellite and street-view matching ranked by CLIP and DINOv2
  • Returns coordinates, error radius, camera heading and evidence image
  • Works in Claude Code, Codex, Cursor, Gemini CLI, OpenCode and Copilot

Install This Skill

Add this skill to your favorite AI agent in a few steps.

Any AI agent

This skill is plain instructions — it works with any assistant that accepts custom instructions or system prompts.

  1. Copy the skill content with the button below.
  2. Paste it into your agent's instruction file or system prompt (for example AGENTS.md, .cursorrules, or a custom instructions field).
  3. Ask the agent to apply the skill whenever the task matches.

Skill Content

Markdown Content

Copy this content and use it with your preferred AI agent

# geo-sleuth

An agent skill that geolocates a photograph and shows its work. Give your agent one or more photos and ask "where was this taken?"; it returns coordinates with an error radius, the camera heading, a satellite evidence image and a graded confidence.

## How it works
1. **Intake** – one script reads EXIF metadata, runs OCR and reverse image search, and logs every clue on a candidate board.
2. **Ranking** – scripts, not the model, score and sort candidate locations. Lookup tables (licence plates, phone codes, driving side, territories) turn clues into hypotheses.
3. **Verification** – OpenStreetMap (Overpass) geometry, elevation-based skyline rendering, sun and shadow maths, CLIP-ranked satellite scans and street-view matching narrow the field. The model only judges among the top few candidates.
4. **Report** – every conclusion must cite the command that ran and the file it produced.

## Install
```bash
npx skills add Oldcircle/geo-sleuth
```
The full SKILL.md, scripts, references and data live in `skills/geo-sleuth/` of the repository: https://github.com/Oldcircle/geo-sleuth

Usage Instructions

Learn how to use this skill with different AI agents.

Generic Instructions

Install with npx skills add Oldcircle/geo-sleuth, then give your agent a photo and say: find where this photo was taken. Python 3.10+ is required for the bundled scripts.

Example Usage

Find where this photo was taken.

Description

geo-sleuth turns any coding agent into a photo-geolocation investigator. Instead of guessing from vibes, it runs a structured workflow: a single intake command extracts EXIF, OCR text and reverse-image-search hits, then a scripted candidate board ranks possible locations so the model only has to judge the top few.

It works even when a photo contains no readable text or landmark. Railway pier spacing becomes a distance ruler, shadows give a bearing and a ridge line becomes a fingerprint that elevation data can be matched against. In the author's showcase run, one stripped-EXIF rice-paddy photo was narrowed from 27,335 railway bridge segments to a single camera position within about 2 m.

The skill follows the open Agent Skills format (SKILL.md plus plain Python scripts), so the same folder works in Claude Code, Codex, Cursor, Gemini CLI, OpenCode and GitHub Copilot. Requires Python 3.10+. MIT licensed. Useful for OSINT researchers, journalists verifying imagery and geography-game fans; every answer includes coordinates, an error radius, camera heading, a satellite evidence image and a graded confidence.

Related Skills

Skill: Trail of Bits Static Analysis

by Trail of Bits

New

CodeQL, Semgrep and SARIF static-analysis toolkit from Trail of Bits: taint tracking, fast pattern scans and merged, deduplicated security findings for coding agents.

Development
5 views

Skill: Playwright CLI

by Microsoft

New

Microsoft's official Playwright skill — drives a real browser from the command line using accessibility snapshots and element refs, and plans, generates and heals Playwright tests.

Development
7 views 1 copies

Skill: Google Maps Platform

by Google Maps Platform

New

Google's official agent skill for writing production Maps Platform code — grounded in freshly fetched docs, with a demo key path that needs no billing account.

Development
9 views
Browse all skills →