feat: Chrome extension + monorepo restructure#10
Open
jiweiyuan wants to merge 14 commits into
Open
Conversation
…e as monorepo Add a Manifest V3 Chrome extension that connects to the REPL server via WebSocket (/ws endpoint) using JSON-RPC 2.0. The extension provides tabs, DOM, page, and cookie control through Chrome Extensions APIs, with Playwriter-inspired visual feedback (ghost cursor, element highlight). Restructure the repo into a workspace monorepo with three symmetric packages: - packages/protocol: shared @browseruse/protocol JSON-RPC types - packages/cli: server, CLI, and CDP session (@browseruse/cli) - extension/: Chrome extension (built with esbuild) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…ull CDP access Replace the offscreen WebSocket bridge with native messaging (chrome.runtime.connectNative) and replace chrome.tabs/scripting/cookies APIs with chrome.debugger for full CDP protocol access (all 56 domains, 652 methods) without requiring --remote-debugging-port=9222. - Add native messaging host bridge (native-host.ts) that relays between Chrome's stdin/stdout protocol and the REPL server WebSocket - Add install script for native host manifest (macOS/Linux) - Add debugger-handler.ts for chrome.debugger attach/detach/sendCommand - Rewrite service worker to use connectNative instead of offscreen doc - Reduce content script to visual feedback only (cursor, highlight, ripple) - Update ws-handler to fall back to extension debugger for session.cdpRaw - Remove offscreen document, page-handlers, network-handlers BREAKING CHANGE: extension no longer uses offscreen WebSocket bridge or chrome.scripting/cookies APIs. All page/DOM/network operations must now go through session.cdpRaw or debugger.sendCommand via chrome.debugger.
Update root package.json scripts, tsconfig references, .gitignore, and fix relative paths in build.ts and install.ts to reflect the new location two levels below the repo root.
Drop chrome.runtime.connectNative / native-host bridge in favor of a direct WebSocket connection to ws://127.0.0.1:9876/ws from the service worker. Remove nativeMessaging permission, delete native-host.ts and native-messaging-host.json template. Rewrite install.ts to clean up legacy native host manifests. Add AUTH_FAILED error code to protocol. BREAKING CHANGE: the extension no longer uses native messaging; the native host bridge and install --extension-id workflow are removed.
Listen on ~/.browseruse/browseruse.sock for NDJSON control protocol, enabling Sarea and scripts to communicate with the REPL without HTTP. Refactor runServer() to accept ServerOptions (silent, controlSocket) and return ServerContext (session, server, startedAt). Add socket cleanup on SIGINT/SIGTERM and /quit.
The root tsconfig.json references packages/protocol but the package had no tsconfig.json, causing tsc --noEmit to fail.
Pre-flight fetch to /health verifies the server is a browseruse REPL before attempting a WebSocket upgrade, preventing 404 errors when a different service or older REPL without /ws occupies the port.
Shows an "Open Sarea" button in the extension popup when the REPL server is not connected, allowing users to launch the Sarea app directly via its io.corespeed.sarea:// URL scheme.
Replace placeholder extension icons with Sarea lion cub icon (16/48/128px). Remove Open Sarea button from popup to keep the UI minimal.
Resolve the browseruse extension from the Sarea app bundle
(Resources/browseruse-extension/) or from the development tree
(packages/extension/dist/). Pass --load-extension and
--disable-extensions-except flags so the extension is loaded
automatically without manual installation.
Also update the release workflow to build and package the
extension as browseruse-extension-{version}.zip alongside
the CLI binaries.
The CLI entry point moved from src/cli.ts to packages/cli/src/cli.ts during the monorepo restructure.
Only use --load-extension for non-system profiles where the store extension isn't installed. For system profiles, skip it to avoid loading a duplicate alongside the Chrome Web Store version. Also remove --disable-extensions-except so user-installed extensions (including the store version) are not disabled.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
/wsendpoint) using JSON-RPC 2.0, providing tabs, DOM, page, and cookie control through Chrome Extensions APIspackages/protocol(shared types),packages/cli(server + CLI),extension/(Chrome extension built with esbuild)Test plan
bun test— 37/37 unit + repl tests pass (no regressions)bun run build:ext— extension builds without errorsextension/dist/in Chrome developer mode → no errorsbrowseruse --start) → extension connects (green dot in popup)tabs.listvia WebSocket → returns current tabsdom.click→ clicks element with visual cursor feedbackpage.screenshot→ returns base64 PNG🤖 Generated with Claude Code