mosaic-apim
AGENTS.md3 starsChanged 12 days agoMITView on GitHub
The README's Screenshots section and the full gallery show people what MOSAIC looks like. Treat those images and their descriptions as documentation that changes with the UI. - When a change alters a pictured page, whether its layout, wording, navigation, or the data it shows, regenerate the affected shots in the same change. Then reread their gallery and README descriptions and correct any that no longer match. Each description is one or two sentences about what the view is and…
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,439 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
The file
We split this AGENTS.md into the rules it gives an agent
5 rules. About 0 are common, 4 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 7
Treat those images and their descriptions as documentation that changes with the UI.
- Distinctive: line 9
When a change alters a pictured page, whether its layout, wording, navigation, or the data it shows, regenerate the affected shots in the …
- Distinctive: line 13
When a change adds a capability worth showing, add a shot for it.
- Very specific: line 15
Keep four things in step: SHOTS in scripts/screenshots/capture.py, the PNGs in docs/images/screenshots/, the gallery and README selection, …
- Distinctive: line 25
If you can't run the capture, for example because the environment has no browser or can't download one, say so in your pull request.
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md 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
Agent instructions
Keep the documentation screenshots current
The README's Screenshots section and the full gallery show people what MOSAIC looks like. Treat those images and their descriptions as documentation that changes with the UI.
- When a change alters a pictured page, whether its layout, wording, navigation, or the data it shows, regenerate the affected shots in the same change. Then reread their gallery and README descriptions and correct any that no longer match. Each description is one or two sentences about what the view is and what it lets someone see or do.
- When a change adds a capability worth showing, add a shot for it. When it renames or removes a pictured page, rename or remove the shot.
- Keep four things in step:
SHOTSinscripts/screenshots/capture.py, the PNGs indocs/images/screenshots/, the gallery and README selection, and the inventory indocs/screenshots.md. That inventory lists which pages are pictured. - Capture only the fictional demo estate the script seeds. Never screenshot a real tenant, subscription, deployment, or person, and never commit an image made any other way.
- Open every regenerated PNG before you commit it. Endpoint URLs, host, resource, and account names, tenant and object IDs, MOSAIC record IDs, and keys must all be blurred. If something slipped through, extend the redaction rules in the script rather than editing the image.
- Debug captures named
*.failed.pngare unredacted. They stay in the script's temporary folder and must never be committed. - If you can't run the capture, for example because the environment has no browser or can't download one, say so in your pull request. Name the shots that need regenerating, so that the documentation isn't left showing an outdated UI without anyone knowing.
docs/screenshots.md has the command, its options, the redaction rules, and how to add a shot.
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.

