mobbin-search
Skill0 starsChanged 20 days agoMITView on GitHub
Search Mobbin's REST API for real iOS or web UI screenshots and visually analyze the downloaded results. Use for UI/UX references, pattern comparisons, design inspiration, or questions about how existing products implement a screen or flow.
- Reads credentials
In this repository11 other files it gives its agents.
- AGENTS.mdAGENTS.md
- app-icon-studioSkill
- disk-cleanupSkill
- localize-app-store-screenshotsSkill
- macos-direct-distributionSkill
- macos-finder-extensionSkill
- smooth-alpha-edgesSkill
- stock-value-analysisSkill
- task-artifactsSkill
- write-ai-docSkill
- xcode-cloud-setupSkill
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 2 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 14
Default to 5 results, deep mode, and optimized images.
- Very specific: line 21
Read the JSON result and inspect every successful local_path with the local image viewer.
- Very specific: line 37
Never print, persist, interpolate into command arguments, or include MOBBIN_API_KEY in generated artifacts.
- Distinctive: line 38
Keep the API key in 1Password.
- Distinctive: line 39
Treat signed image URLs as temporary credentials: use them for immediate inspection and do not publish them.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Mobbin Search
Ground design advice in screenshots returned by Mobbin rather than assumptions.
Search
- Choose a focused natural-language query describing one screen or intent. Do not include the platform in the query.
- Infer
iosfor SwiftUI, React Native, or mobile work andwebfor browser products. Ask only when the platform is genuinely unclear. - Default to 5 results,
deepmode, and optimized images. Use up to 15 results when the user explicitly wants variety or comparison. Briefly announce the query and result count, then run:
<skill-directory>/scripts/run-search --platform <ios|web> --limit <count> "<query>"- Read the JSON result and inspect every successful
local_pathwith the local image viewer. Never describe a screen from metadata alone.
The launcher accepts a direct MOBBIN_API_KEY environment variable. Otherwise it resolves the secret through 1Password using mobbin.env. If configuration is missing, explain the one-time setup from the launcher's error without asking the user to paste a secret into chat.
Respond
Ground observations in visible details such as hierarchy, copy, controls, spacing, color, and element count. Link each cited example using its mobbin_url.
- Give a direct answer when 1–3 screens and a short explanation resolve the question.
- For broad comparisons or inspiration sets, first offer a side-by-side HTML evidence board, a user-defined layout, or a text summary. Build the board only after the user chooses.
- Save any evidence board under
.mobbin/, keep it self-contained, and link every screen back to Mobbin.
Downloaded images and results.json are temporary research artifacts under .mobbin/; do not commit them.
Security
- Never print, persist, interpolate into command arguments, or include
MOBBIN_API_KEYin generated artifacts. - Keep the API key in 1Password. The committed example contains only a secret reference.
- Treat signed image URLs as temporary credentials: use them for immediate inspection and do not publish them.
- On HTTP 401 or 403, report that the key or Mobbin plan needs attention; do not retry repeatedly.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

