我去，AI界的福尔摩斯Skill项目
只需给一张照片，他们就能定位拍摄地点，返回相机位置、拍摄方向及卫星图像作为证据。
Oldcircle/geo-sleuth t.co
3 likes · 259 views
仓库：Oldcircle/geo-sleuth
安装：npx skills add Oldcircle/geo-sleuth
1 like · 11 views
An agent skill that finds where a photo was taken — OpenStreetMap geometry, elevation skylines, satellite imagery and street view — and shows its work. Works with Claude Code, Codex, Cursor, Gemini CLI, OpenCode and GitHub Copilot.
Oldcircle/geo-sleuth
14 likes · 2 reposts · 1,243 views
You can use Claude Code to find where a street photo was taken, within three meters, in under an hour. A free skill called geo-sleuth did it on a Vancouver cherry tree photo with no GPS data.
Get it here: Oldcircle/geo-sleuth t.co
6 likes · 1 repost · 210 views
geo-sleuth: An agent skill that finds where a photo was taken (2.3k stars)
What it does
All text below is quoted directly from the project README.
>An agent skill that finds where a photo was taken — and shows its work.
>No text. No plates. No landmarks. One bridge, one mountain. Located to within 2 m.
Why geo-sleuth:
- One photo, one sentence. Give your agent a photo and say find where this photo was taken. You get back the camera position, the direction it was facing, and a satellite evidence image.
- It works when there is nothing to read. No sign, no plate, no landmark: OpenStreetMap geometry, elevation data, satellite tiles and street view carry the search on their own.
- Geometry instead of guesswork. Pier spacing becomes a distance ruler, shadows become a bearing, a ridge line becomes a fingerprint that elevation data can be matched against.
- Every claim points at a file. A conclusion has to name the command that ran in the session and the file it produced. Population and fame are not evidence.
- Scripts rank, the model judges. Twenty single-purpose scripts search, score and sort; the model only picks among the top few.
- One skill, every agent. A standard Agent Skill — SKILL.md plus plain Python scripts — so the same folder runs in Claude Code, Codex, Cursor, Gemini CLI, OpenCode and GitHub Copilot.
- Answers carry an error radius. Coordinates ± radius, the camera heading, an evidence image and a graded confidence.
>Pick your agents when prompted. Then hand your agent a photo and say: find where this photo was taken. That is the whole interface. The agent reads SKILL.md, runs the scripts, and comes back with the camera position, the direction it was facing and a satellite evidence image.
>geo-sleuth is a standard Agent Skill: one folder holding SKILL.md, scripts/, references/ and data/. Install it with the skills CLI, or copy the folder yourself.
Repository
Oldcircle/geo-sleuth
Repository stats
|Metric|Value|
|:-|:-|
|Stars|2,263|
|Forks|284|
|Last release|None|
|Languages|Python 100%|
|Contributors|5|
|AI contributor accounts (Claude / Codex / Cursor / Copilot)|claude (10 commits)|
|AI co-authored commits|Claude: 9|
Main contributor
|Field|Value|
|:-|:-|
|Name|yuanbo|
|GitHub|@Oldcircle|
|Commits to this repo|18 (top contributor)|
|Company / Location|Central China Normal University / China|
|Followers|48|
|Public repos|94|
|On GitHub since|2020-10-09|
324 points · 19 comments
geo-sleuth
Just recently cam across this tool: Oldcircle/geo-sleuth
I tested it with various images where I knew the location. Easy images were almost always located quickly, the harder ones (e.g., no signs/flags/number plates/famous buildings) were hit or miss.
Most impressive hit were when I gave it three pictures I took at a Danish beach. You could only see the sea, a beach house, dunes, and a couple of offshore windmills. It managed to locate the beach because it guessed Denmark from the dunes + house, then got it correct that I was facing west and because there are only two offshore wind parks with a specific arrangement of windmills. Generally impressive.
Do you have experience with this tool? Do you know other tools like it?
15 points · 5 comments
I came across geo-sleuth tonight - an agent skill that pinpoints where a photo was taken with zero text or landmarks in frame. it reads bridge piers as a ruler, a mountain ridge as a fingerprint matched to elevation data. nailed its demo to within 2m.
Oldcircle/geo-sleuth t.co
2 likes · 119 views
このスキルで写真から撮影場所を特定できるらしいが果たして
Oldcircle/geo-sleuth
Mentions
我去，AI界的福尔摩斯Skill项目 只需给一张照片，他们就能定位拍摄地点，返回相机位置、拍摄方向及卫星图像作为证据。 Oldcircle/geo-sleuth t.co
仓库：Oldcircle/geo-sleuth 安装：npx skills add Oldcircle/geo-sleuth
An agent skill that finds where a photo was taken — OpenStreetMap geometry, elevation skylines, satellite imagery and street view — and shows its work. Works with Claude Code, Codex, Cursor, Gemini CLI, OpenCode and GitHub Copilot. Oldcircle/geo-sleuth
You can use Claude Code to find where a street photo was taken, within three meters, in under an hour. A free skill called geo-sleuth did it on a Vancouver cherry tree photo with no GPS data. Get it here: Oldcircle/geo-sleuth t.co
geo-sleuth: An agent skill that finds where a photo was taken (2.3k stars) What it does All text below is quoted directly from the project README. >An agent skill that finds where a photo was taken — and shows its work. >No text. No plates. No landmarks. One bridge, one mountain. Located to within 2 m. Why geo-sleuth: - One photo, one sentence. Give your agent a photo and say find where this photo was taken. You get back the camera position, the direction it was facing, and a satellite evidence image. - It works when there is nothing to read. No sign, no plate, no landmark: OpenStreetMap geometry, elevation data, satellite tiles and street view carry the search on their own. - Geometry instead of guesswork. Pier spacing becomes a distance ruler, shadows become a bearing, a ridge line becomes a fingerprint that elevation data can be matched against. - Every claim points at a file. A conclusion has to name the command that ran in the session and the file it produced. Population and fame are not evidence. - Scripts rank, the model judges. Twenty single-purpose scripts search, score and sort; the model only picks among the top few. - One skill, every agent. A standard Agent Skill — SKILL.md plus plain Python scripts — so the same folder runs in Claude Code, Codex, Cursor, Gemini CLI, OpenCode and GitHub Copilot. - Answers carry an error radius. Coordinates ± radius, the camera heading, an evidence image and a graded confidence. >Pick your agents when prompted. Then hand your agent a photo and say: find where this photo was taken. That is the whole interface. The agent reads SKILL.md, runs the scripts, and comes back with the camera position, the direction it was facing and a satellite evidence image. >geo-sleuth is a standard Agent Skill: one folder holding SKILL.md, scripts/, references/ and data/. Install it with the skills CLI, or copy the folder yourself. Repository Oldcircle/geo-sleuth Repository stats |Metric|Value| |:-|:-| |Stars|2,263| |Forks|284| |Last release|None| |Languages|Python 100%| |Contributors|5| |AI contributor accounts (Claude / Codex / Cursor / Copilot)|claude (10 commits)| |AI co-authored commits|Claude: 9| Main contributor |Field|Value| |:-|:-| |Name|yuanbo| |GitHub|@Oldcircle| |Commits to this repo|18 (top contributor)| |Company / Location|Central China Normal University / China| |Followers|48| |Public repos|94| |On GitHub since|2020-10-09|
geo-sleuth Just recently cam across this tool: Oldcircle/geo-sleuth I tested it with various images where I knew the location. Easy images were almost always located quickly, the harder ones (e.g., no signs/flags/number plates/famous buildings) were hit or miss. Most impressive hit were when I gave it three pictures I took at a Danish beach. You could only see the sea, a beach house, dunes, and a couple of offshore windmills. It managed to locate the beach because it guessed Denmark from the dunes + house, then got it correct that I was facing west and because there are only two offshore wind parks with a specific arrangement of windmills. Generally impressive. Do you have experience with this tool? Do you know other tools like it?
I came across geo-sleuth tonight - an agent skill that pinpoints where a photo was taken with zero text or landmarks in frame. it reads bridge piers as a ruler, a mountain ridge as a fingerprint matched to elevation data. nailed its demo to within 2m. Oldcircle/geo-sleuth t.co
このスキルで写真から撮影場所を特定できるらしいが果たして Oldcircle/geo-sleuth