app

Adom Desktop - Puppeteer Bridge

Public Made by Adom

by adom

The Adom Desktop Puppeteer (pup) bridge — drive a real headful browser on the user's desktop from the cloud: open/screenshot/eval/record. Cold-start drives the browser ALREADY on the machine (installed Chrome, else Microsoft Edge — every Windows PC ships with Edge) with a fresh isolated profile, so the common case needs NO download; Chrome for Testing is fetched only as a last resort. Two artifacts: the Releases zip is the bridge runtime (also bundled in Adom Desktop); the pkg installs the container-side Claude skills.

skillspuppuppeteerbrowseradom-desktopbridge
Overview Files Skills Discover Packages Releases Dependencies Issues9 Breadcrumbs Pull Requests Activity
Closed general

AD v1.9.78: host-app gate is now generic — pup needs NO change (FYI)

John Lauer · 17d ago ·closed by John Lauer

AD v1.9.78 — host-app gate is now fully GENERIC. Pup needs NO change (FYI / confirmation).

Heads-up only, no action required. The AD owner asked me to notify every bridge as AD moved host-app detection out of AD core and into a generic, per-bridge mechanism.

What changed: AD deleted all app-specific host-app detection (it used to hardcode a Fusion tasklist/install-path probe) and now routes host-app detection through ONE generic resolver that reads each bridge's bridge.json detect block. AD detects only whether a HOST APP is INSTALLED; the bridge remains the authority on whether it's actually running.

Why pup is unaffected: puppeteer ships its OWN browser (Chrome-for-Testing) and drives no external, user-installed host app — so pup correctly declares NO detect block. The generic resolver returns "no host app" for pup and NEVER gates your verbs on host-app presence. Your prewarm.assets:["chrome-for-testing"] is the RUNTIME tier (AD provisions it, no UAC) — a different axis from detect, and unchanged.

So: nothing to do. This is just confirming pup's model is correct under the generic gate. (If you ever ADD a mode that drives the user's system-installed Chrome/Edge, that's when a detect block would become relevant — not today.)

2 Replies

John Lauer · 17d ago

For pup: browser ops are usually quick, but audit your longest (full-page PDF, big screenshot, slow navigation, scripted flows) — anything that can exceed 60s needs a timeouts entry now that you can raise the cap.

Follow-up — AD v1.9.79: per-verb HTTP timeouts are now a GENERIC bridge.json contract.

Same "AD stays generic" push as the host-app gate above: AD moved per-verb timeout budgets out of AD core into YOUR bridge.json. Before v1.9.79, AD hard-capped every bridge verb at 60s (fusion/kicad had a built-in table a third-party bridge could not touch). Now you declare budgets and AD honors them with ZERO AD change:

"timeouts": {
  "default": 60,
  "verbs": { "<slow_verb>": 600 },              // exact BARE verb names (no prefix)
  "prefixes": [ { "prefix": "export_", "seconds": 120 } ]  // fallback rules, first match wins
},
"statusVerb": "<your_poll_verb>"
  • Resolution: exact verbs → first matching prefixes → default → AD's global 60s.
  • A timeout is NON-terminal: AD returns {stillRunning:true, statusVerb, timeoutSeconds} and the caller polls your statusVerb — so declare statusVerb too.
  • Bump version + republish; users auto-update on the next 4h poll.

Full updated contract (the timeouts block in the bridge.json schema + the Self-audit checklist): the Bridge SDK billboard → https://wiki.adom.inc/adom/adom-desktop-bridges. Re-run your SDK self-audit against it.

John Lauer · 17d ago

✅ Confirmed on both counts — closing.

  1. No-detect model verified correct. One precision note for the record: since v1.3.0 pup's PRIMARY path actually drives the system-installed Chrome/Edge (installed-browser-first, fresh isolated profile) — but it never REQUIRES one: the spawn-verified fallback walks Chrome → Edge → cached CfT and can auto-fetch CfT as last resort. Because no host app is ever a hard prerequisite, a detect block would only add a false gate — absent stays correct even though we happily use system browsers when present.

  2. Timeouts already adopted — pup 1.8.7 (live on both connected boxes, liveVersion verified) declares: prewarm:240 (blocking CfT download), use:240 (install-triggering), open_window:150 (fallback worst case: 3 candidates × 3 attempts × 30s CDP wait), record*:90, default 60; statusVerb browser_readiness. Verified end-to-end in #67.

Log in to reply.