app-screenshots
Skill1 starsChanged 2 days agoMITView on GitHub
Capture every page of a running Next.js (or any web) app in light and dark, desktop and phone, and every locale, from fresh data, and check each capture for blank pages, error overlays, horizontal scroll on phone, fonts or images that did not load, the wrong theme and the wrong text direction. Use when asked for screenshots of the whole app, a screenshot grid, release or listing screenshots, or a visual check of every route.
- Installs packages
7 other projects use a version of this.
Adopted by
Most recently taken up by
- matthiasn/lottitheir own version, first seen Nov 8, 2026
- itsryanarmani/skillstheir own version, first seen Nov 8, 2026
- buddyh/agent-skillstheir own version, first seen Nov 8, 2026
- pooriaarab/skillstheir own version, first seen Nov 7, 2026
- mikezamayias/agent-skillstheir own version, first seen Oct 11, 2026
- Foundry376/Mailspringtheir own version, first seen Oct 8, 2026
Other versions7 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 7
- Foundry376/Mailspring★ 18ksame name
- matthiasn/lotti★ 1.2ksame name
- buddyh/agent-skills★ 5same name
- itsryanarmani/skills★ 1same name
- mikezamayias/agent-skills★ 1same name
- pooriaarab/skills★ 1same name
- edisonjoao1/app-screenshots-skill★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository10 other files it gives its agents.
- add-admin-dashboardSkill
- design-rulesSkill
- docs-checkSkill
- i18n-keys-checkSkill
- nextjs-claude-mdSkill
- nextjs-hooksSkill
- production-checkSkill
- rtl-checkSkill
- template-rebrandSkill
- vps-deploySkill
The file
We split this skill into the rules it gives an agent
3 rules. 0 are common, 0 are distinctive, and 3 are very specific to this project.
0
Common
Most of the 5 edited versions we read keep these.
0
Distinctive
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 19
Check: every page.* file is in routes or written down as skipped, with the reason.
- Very specific: line 23
Run the project's reset and seed commands (for example npm run db:reset && npm run db:seed).
- Very specific: line 86
Add screenshots/ to .gitignore: screenshots/.auth/ holds signed-in browser state.
“Common” is measured: kept by at least 6 in 10 of the 5 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
App screenshots
Writes screenshots/<locale>/<theme>/<device>/<route>.png and screenshots/report.md.
A capture that fails a check is listed with the reason, and the run exits 1.
1. List the routes
Find every page: find app src/app -name 'page.*' 2>/dev/null | sort. Write
screenshots.config.json at the project root, starting from config.example.json beside
this file. Put {locale} where the app has a locale segment, give each dynamic segment
([slug]) a real value from the sample data, and give pages behind sign-in "as": "<login>".
An app with one fixed look and no light or dark mode uses "themes": ["default"].
Check: every page.* file is in routes or written down as skipped, with the reason.
2. Start from fresh data
Run the project's reset and seed commands (for example npm run db:reset && npm run db:seed).
An app on browser-stored sample data needs nothing here: every capture opens a new browser
context from the state saved right after sign-in.
Check: the seed exits 0, and one record it creates is visible in the API or database.
3. Build and start on a free port
curl -s -o /dev/null http://localhost:3043 && echo "3043 is taken, pick another"
npm run build
node node_modules/next/dist/bin/next start -p 3043 & echo $! > .screenshots.pidA production build: next dev compiles each route on its first visit, so timings drift and
dev-only overlays appear. Set the same port in baseUrl.
Check: curl -s -o /dev/null -w '%{http_code}' http://localhost:3043/ prints 200 (or the
307/308 your root redirect sends).
4. Make Playwright available
node -e "console.log(require.resolve('playwright'))" in the project. If that fails, install
it outside the project and point the script at it:
mkdir -p /tmp/pw && cd /tmp/pw && npm init -y && npm i playwright && npx playwright install chromiumCheck: the resolve prints a path, or /tmp/pw/node_modules/playwright exists.
5. Capture
PLAYWRIGHT_DIR=/tmp/pw node ${CLAUDE_SKILL_DIR}/scripts/capture.mjs screenshots.config.json(Drop PLAYWRIGHT_DIR when the project has Playwright.) Each capture gets its own context
with the theme set as prefers-color-scheme, scrolls to the end so lazy images load, waits
for fonts and images, runs the checks, hides the hide selectors, grows the viewport to the
page height and takes the picture. Light and dark of the same page are then compared byte
for byte.
Check: the last line says 0 failed and the exit code is 0. Otherwise read
screenshots/report.md, fix the page or the config, and rerun only those captures with
--only <part of the id>, for example --only ar/dark/phone.
6. Look at every capture
Open each PNG with the Read tool. The checks cannot see cut-off text, overlapping cards, icons that did not mirror in right-to-left, an empty chart, placeholder copy or a wrong picture. Write one line per capture, naming the file and anything wrong.
Check: the notes have a line for every PNG, and each problem is fixed or reported.
7. Stop the server you started
kill "$(cat .screenshots.pid)" && rm .screenshots.pidCheck: curl -s http://localhost:3043/ can no longer connect.
Add screenshots/ to .gitignore: screenshots/.auth/ holds signed-in browser state.
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.

