browse
Skill0 starsChanged 5 months agoMITView on GitHub
Interactive browser QA using xd://browser or (optionally) the gstack browse binary for headed mode. Navigate pages, snapshot accessibility trees, interact via element refs, capture screenshots, verify page state, and hand off for CAPTCHA. Use when you need to test a page live, debug a UI bug, verify a deployment, or dogfood a user flow.
53 other projects use a version of this.
Adopted by
Most recently taken up by
- lohi-ai/babysittheir own version, first seen Nov 8, 2026
- MildlyMeticulous/shortlisttheir own version, first seen Nov 8, 2026
- timurgaleev/vibestacktheir own version, first seen Nov 8, 2026
- WordPress/gutenbergtheir own version, first seen Nov 1, 2026
- AhmedMerza/claude-skillstheir own version, first seen Nov 1, 2026
- AhmedMerza/agent-skillstheir own version, first seen Nov 1, 2026
Used with
Projects with it also use setup-gbrain, open-gstack-browser and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
ShreyRaviSkillsetup-gbrain
Set up gbrain for this coding agent: install the CLI, initialize a local PGLite or Supabase brain, register MCP, capture per-remote trust policy. One command from zero to "gbrain is running, and this agent can call it." Use when: "setup gbrain", "connect gbrain", "start gbrain", "install gbrain", "configure gbrain for this machine". (gstack)
Also in 9 projects that use this
seek888Skillopen-gstack-browser
Launch GStack Browser — AI-controlled Chromium with the sidebar extension baked in. Opens a visible browser window where you can watch every action in real time. The sidebar shows a live activity feed and chat. Anti-bot stealth built in. Use when asked to "open gstack browser", "launch browser", "connect chrome", "open chrome", "real browser", "launch chrome", "side panel", or "control my browser". Voice triggers (speech-to-text aliases): "show me the browser".
Also in 11 projects that use this
Saqlain-811Skillios-sync
Regenerate the iOS debug bridge against the latest upstream gstack templates. (gstack)
Also in 6 projects that use this
ayudaySkillsetup-deploy
Configure deployment settings for /land-and-deploy. Detects your deploy platform (Fly.io, Render, Vercel, Netlify, Heroku, GitHub Actions, custom), production URL, health check endpoints, and deploy status commands. Writes the configuration to CLAUDE.md so all future deploys are automatic. Use when: "setup deploy", "configure deployment", "set up land-and-deploy", "how do I deploy with gstack", "add deploy config".
Also in 19 projects that use this
ShreyRaviSkilllanding-report
Read-only queue dashboard for workspace-aware ship. Shows which VERSION slots are currently claimed by open PRs, which sibling Conductor workspaces have WIP work likely to ship soon, and what slot /ship would pick next. No mutations — just a snapshot. Use when asked to "landing report", "what's in the queue", "show me open PRs", or "which version do I claim next". (gstack)
Also in 9 projects that use this
Saqlain-811Skillios-clean
Remove the DebugBridge SPM package and all #if DEBUG wiring from an iOS app. (gstack)
Also in 5 projects that use this
- +3 more from ShreyRavi
- +1 more from Saqlain-811
Other versions53 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 53
- garrytan/gstack★ 141ksame name
- WordPress/gutenberg★ 12ksame name
- SeifBenayed/cloclo★ 114same name
- mr-daedalium/ostack-saas★ 108same name
- no-session/pstack★ 96same name
- lucas-flatwhite/gstack-ko★ 45same name
- shipshitdev/skills★ 39same name
- WillNigri/Agentic-Tool-Optimization★ 34same name
- Scino/fstack★ 18same name
- ulpi-io/browse★ 15same name
43 more not listed
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 repository38 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- api-contract-validationSkill
- architecture-scannerSkill
- auto-approve-policySkill
- auto-learningSkill
- backend-patternsSkill
- branch-syncSkill
- browser-testSkill
- build-checkSkill
- cache-refreshSkill
- codebase-explainerSkill
- coding-standardsSkill
- commit-styleSkill
- context-switchingSkill
- continuous-learningSkill
- db-preflightSkill
- env-preflightSkill
- eval-harnessSkill
- feature-completenessSkill
- file-indexSkill
- frontend-patternsSkill
- function-extractSkill
- idev-initSkill
- import-graphSkill
- lessons-learnedSkill
- post-creation-verifySkill
- pre-ship-verifySkill
- project-mapSkill
- self-reviewSkill
- session-resumeSkill
- skill-benchmarkSkill
- skilloptSkill
- smart-contextSkill
- strategic-compactSkill
- task-journalSkill
- tdd-workflowSkill
- test-coverageSkill
- test-mapSkill
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 45 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 24
If GSTACK_BROWSE=none, all interactions use xd://browser (headless).
- Very specific: line 25
If GSTACK_BROWSE is set, $B works for all gstack browse commands.
- Very specific: line 196
After they solve it: $B resume to re-snapshot and continue.
“Common” is measured: kept by at least 6 in 10 of the 45 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Browse — Interactive Browser QA
Two backends: xd://browser (runtime built-in, zero install, headless only)
or gstack browse (97KB binary, headed mode + handoff for CAPTCHA).
The skill auto-detects which is available.
Setup (run first)
# Detect gstack browse binary
B=""
[ -x /home/theo/.claude/skills/gstack/browse/dist/browse ] && B="/home/theo/.claude/skills/gstack/browse/dist/browse"
[ -x /home/theo/.codex/skills/gstack/browse/dist/browse ] && B="/home/theo/.codex/skills/gstack/browse/dist/browse"
[ -x "$(git rev-parse --show-toplevel 2>/dev/null)/.agents/skills/gstack/browse/dist/browse" ] && B="$(git rev-parse --show-toplevel 2>/dev/null)/.agents/skills/gstack/browse/dist/browse"
export B
if [ -n "$B" ]; then echo "GSTACK_BROWSE=$B"; else echo "GSTACK_BROWSE=none (using xd://browser)"; fiIf GSTACK_BROWSE=none, all interactions use xd://browser (headless).
If GSTACK_BROWSE is set, $B works for all gstack browse commands.
Quick Reference
# xd://browser (always available — headless only)
xd://browser { "action": "open", "url": "https://app.com", "name": "main" }
run code: tab.ariaSnapshot() # ARIA YAML with [ref=eN]
run code: tab.fill('aria-ref=e2', 'Alice')
run code: tab.click('aria-ref=e5')
# gstack browse (headed + handoff — when available)
$B goto https://app.com
$B snapshot -i # @e1, @e2... refs
$B fill @e2 "Alice"
$B click @e5
$B handoff "solve CAPTCHA" # visible browser for user
$B resume # continue after handoffCore Workflow
- Detect backend via the Setup block above
- Open tab —
xd://browser openor$B goto - Snapshot —
tab.ariaSnapshot()or$B snapshot -i - Interact — fill, click, select by @ref
- Verify — check text, URL, console errors
- Handoff if stuck —
$B handofffor CAPTCHA/complex auth - Screenshot — capture evidence
- Close — close tab or
$B stop
Patterns using xd://browser (always available)
Pattern 1: Basic page verification
xd://browser { "action": "open", "url": "https://app.com/login" }
xd://browser { "action": "run", "name": "main", "code": "tab.observe()" }
xd://browser { "action": "run", "name": "main",
"code": "const errs=[]; page.on('console',m=>m.type()==='error'&&errs.push(m.text())); page.on('requestfailed',r=>errs.push('NET: '+r.url())); await tab.goto('https://app.com/login'); errs" }Pattern 2: Snapshot + interact by @ref
xd://browser { "action": "run", "name": "main", "code": "tab.ariaSnapshot()" }
xd://browser { "action": "run", "name": "main",
"code": "await tab.fill('aria-ref=e2', 'user@test.com'); await tab.click('aria-ref=e5'); tab.screenshot()" }
xd://browser { "action": "run", "name": "main", "code": "tab.observe()" }Refs invalidate on navigation — re-observe after every goto.
Pattern 3: Form fill and submit
xd://browser { "action": "run", "name": "main", "code": "
await tab.fill('input[type=text]', 'Alice');
await tab.select('select', 'green');
await tab.click('button');
await tab.waitForSelector('#result');
tab.extract('#result')" }Pattern 4: Screenshot evidence
xd://browser { "action": "run", "name": "main",
"code": "tab.screenshot()" }
xd://browser { "action": "run", "name": "main",
"code": "tab.screenshot({ selector: '.dashboard' })" }Pattern 5: Dialog handling
xd://browser { "action": "open", "url": "...", "dialogs": "accept", "name": "main" }
xd://browser { "action": "run", "name": "main", "code": "
page.on('dialog', d => d.accept());
await tab.click('#delete-btn');" }Pattern 6: Multi-tab testing
xd://browser { "action": "open", "url": "https://app.com/a", "name": "tab1" }
xd://browser { "action": "open", "url": "https://app.com/b", "name": "tab2" }
xd://browser { "action": "run", "name": "tab1", "code": "tab.observe()" }
xd://browser { "action": "close", "name": "tab1" }Pattern 7: Viewport/responsive
xd://browser { "action": "open", "url": "...",
"viewport": { "width": 375, "height": 812 }, "name": "mobile" }Pattern 8: Custom JS evaluation
xd://browser { "action": "run", "name": "main", "code": "
const result = await tab.evaluate(() => ({
title: document.title, url: location.href
}));
result" }Pattern 9: Wait conditions
xd://browser { "action": "run", "name": "main", "code": "
await tab.waitForSelector('.toast-success', { timeout: 5000 }); 'found it'" }
xd://browser { "action": "run", "name": "main", "code": "
await tab.click('button');
await tab.waitForUrl('**/dashboard');
tab.screenshot()" }Pattern 10: File upload
xd://browser { "action": "run", "name": "main",
"code": "await tab.uploadFile('input[type=file]', '/path/to/file.pdf'); 'uploaded'" }Pattern 11: Drag and drop
xd://browser { "action": "run", "name": "main",
"code": "await tab.drag('#source', '#target'); 'dragged'" }Pattern 12: Cookie and storage manipulation
xd://browser { "action": "run", "name": "main",
"code": "tab.evaluate(() => { document.cookie='session=abc123;path=/';"+
"localStorage.setItem('pref','dark');"+
"return {cookies:document.cookie,storage:{...localStorage}} })" }Pattern 13: Wait for network response
xd://browser { "action": "run", "name": "main", "code": "
const [resp] = await Promise.all([
tab.waitForResponse(r => r.url().includes('/api/submit')),
tab.click('#submit-btn')
]);
{ status: resp.status(), url: resp.url(), ok: resp.ok() }" }Pattern 14: Scroll element into view
xd://browser { "action": "run", "name": "main",
"code": "await tab.scrollIntoView('#footer'); 'scrolled'" }Patterns using gstack browse (when available)
These require $B to be set (from the Setup step above).
Pattern G1: Quick snapshot + interact
$B goto https://app.com/login
$B snapshot -i # @e1 = email, @e2 = password, @e3 = submit
$B fill @e2 "user@test.com"
$B fill @e3 "pass123"
$B click @e4
$B snapshot -i # verify post-login statePattern G2: Snapshot diff (before/after)
$B snapshot -D # baseline stored
$B click @e3 # do something
$B snapshot -D # unified diff against baselinePattern G3: Handoff for CAPTCHA or complex auth
$B handoff "Stuck on CAPTCHA at login page"Then present an AskUserQuestion telling the user a browser is open.
After they solve it: $B resume to re-snapshot and continue.
Pattern G4: Headed browser with extension
$B connect # launches visible Chromium with sidebar
$B disconnect # back to headless modePattern G5: Full QA workflow (gstack)
$B goto https://app.com/login
$B snapshot -i -a -o /tmp/annotated.png # annotated screenshot
$B console --errors # check JS errors
$B fill @e2 "user@test.com"
$B fill @e3 "pass123"
$B click @e4
$B snapshot -D # diff post-login state
$B is visible ".dashboard"
$B screenshot /tmp/dashboard.pngPattern G6: Multi-step with browser state
$B state save login_state # save cookies + URL
$B state load login_state # restore laterPattern G7: Page cleanup and responsive
$B cleanup --all # remove ads, cookies, sticky
$B viewport 375x812 # mobile
$B screenshot /tmp/mobile.png
$B viewport 1280x720 # desktop
$B screenshot /tmp/desktop.pngAdvanced patterns (xd://browser)
Load the on-demand reference for snapshot diff, annotated screenshots, CSS inspection, page cleanup, and URL content comparison:
${CLAUDE_PLUGIN_ROOT}/skills/browse/references/advanced-patterns.mdReference: gstack browse → idev equivalent
| gstack browse | idev equivalent |
$B goto url | open with url (xd://browser) or $B goto (gstack) |
$B snapshot -i | tab.ariaSnapshot() / $B snapshot -i |
$B fill @e2 "v" | tab.fill('aria-ref=e2', 'v') / $B fill @e2 "v" |
$B click @e3 | tab.click('aria-ref=e3') / $B click @e3 |
$B screenshot | tab.screenshot() / $B screenshot |
$B console | page.on('console',...) in run / $B console |
$B handoff "msg" | — (gstack binary only) |
$B connect | — (gstack binary only) |
$B state save/load | — (gstack binary only) |
Gotchas
tab.filldoes NOT work on<select>— usetab.selectinstead.- Navigation invalidates refs for both backends — re-snapshot after
every
goto. $B snapshot -iis an accessibility tree;$B snapshotalone opens GNOME Screenshot on some systems. Always pass-ior-c.- Handoff requires a user to be present — don't use it in unattended runs.
- xd://browser screenshot path is returned, not passed.
- Stalled actions fail fast — don't add sleep/retry loops.
When to use which
| Need | Tool |
| Quick headless verification | browse (xd://browser backend) |
| CAPTCHA / complex auth | browse (gstack browse backend with handoff) |
| Visible browser session | browse (gstack backend, $B connect) |
| Durable E2E test script | browser-test skill |
| Static content read | read tool (no browser needed) |
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.

