rapp-copilot-in-chrome
Drive the user's real, logged-in Chrome browser from GitHub Copilot CLI — navigate, click, type, screenshot, read the accessibility tree or page text, run JavaScript, and inspect console and network traffic. Use when the user asks to open or browse a site, click through a web UI, fill a form, scrape or read a page, debug a web app in the browser, take a screenshot of a page, or automate anything in Chrome. Also use for "rapp copilot in chrome", "browser automation", "drive chrome", "open a tab", "check the console", or to install, verify, or diagnose the bridge itself.
- License 许可
- MIT
- 最近更新
- 2026-11-10
- Stars
- ★ 1
- Forks
- 0
01 前言 Front Matter
- name
- rapp-copilot-in-chrome
- description
- Drive the user's real, logged-in Chrome browser from GitHub Copilot CLI — navigate, click, type, screenshot, read the accessibility tree or page text, run JavaScript, and inspect console and network traffic. Use when the user asks to open or browse a site, click through a web UI, fill a form, scrape or read a page, debug a web app in the browser, take a screenshot of a page, or automate anything in Chrome. Also use for "rapp copilot in chrome", "browser automation", "drive chrome", "open a tab", "check the console", or to install, verify, or diagnose the bridge itself.
02 资源文件 51
- 其他.gitignore43 B
- 参考文档.rapp/member.md1.2 KB
- 其他bin/rapp-copilot-in-chrome1.2 KB
- 脚本bridge.py17.4 KB
- 脚本build_voice_twin_egg.py1.8 KB
- 其他com.rapp.digital-understudy.plist.template1.2 KB
- 其他com.rapp.voice-assistant.plist.template1.3 KB
- 参考文档DIGITAL_UNDERSTUDY.md3.4 KB
- 脚本digital_understudy.py57.0 KB
- 脚本discord_transport.py17.7 KB
- 参考文档docs/tools.json31.0 KB
- 脚本extension/background.js14.4 KB
- 参考文档extension/manifest.json514 B
- 参考文档extension/popup.html1.6 KB
- 脚本extension/popup.js1.4 KB
- 脚本gvoice.py34.8 KB
- 脚本imessage_transport.py47.1 KB
- 脚本install_local.py23.0 KB
- 脚本install_understudy.py10.7 KB
- 脚本install-local.sh593 B
- 脚本install.sh1.7 KB
- 许可证LICENSE1.0 KB
- 脚本messaging_transport.py19.9 KB
- 脚本rapp_copilot_in_chrome_agent.py20.3 KB
- 脚本rapp1.py61.7 KB
- 脚本rappter_chrome_mcp.py13.2 KB
- 其他rappter-voice-assistant.service.template326 B
- 参考文档README.md22.4 KB
- 参考文档supervisor/package-lock.json41.1 KB
- 参考文档supervisor/package.json399 B
- 脚本supervisor/server.js7.7 KB
- 脚本test_bridge.py11.2 KB
- 脚本test_digital_understudy.py18.8 KB
- 脚本test_discord_transport.py10.7 KB
- 脚本test_gvoice.py7.9 KB
- 脚本test_imessage_transport.py16.5 KB
- 脚本test_install_local.py10.0 KB
- 脚本test_install_understudy.py8.7 KB
- 脚本test_mcp.py3.5 KB
- 脚本test_messaging_transport.py14.9 KB
- 脚本test_voice_assistant.py25.5 KB
- 脚本test_voice_twin.py11.9 KB
- 脚本test_whatsapp_transport.py21.1 KB
- 参考文档UNIVERSAL_MESSAGING.md4.8 KB
- 脚本universal_messaging.py8.9 KB
- 脚本voice_assistant.py31.7 KB
- 脚本voice_twin_agent.py7.3 KB
- 参考文档VOICE_TWIN_CONFORMANCE.json2.1 KB
- 参考文档voice_twin_soul.md1.1 KB
- 脚本voice_twin.py32.3 KB
- 脚本whatsapp_transport.py28.8 KB
03 SKILL.md 正文
rapp-copilot-in-chrome
Browser automation against the browser you actually use — real profile, live cookies, authenticated sessions. Not a throwaway headless instance.
Architecture
Copilot CLI (MCP client)
-> ~/.copilot/bin/rapp-copilot-in-chrome (launcher shim)
-> rapp-chrome-supervisor/server.js (resilience proxy -- retries/restarts on hang)
-> claude --claude-in-chrome-mcp (self-contained stdio MCP server)
-> native host com.anthropic.claude_code_browser_extension
-> Chrome extension fcoeoabgfenejglbffodgkkbkcdhcgfn
-> live tabs
The MCP server is self-contained: it does not require a Claude Code session to be running. It needs only the Claude binary (which hosts the bridge) plus the Chrome extension installed and connected. That is the whole trick — the bridge was always a plain stdio MCP server, so any MCP client can drive it.
The supervisor step forwards every call unchanged and only intervenes on a hang/timeout-looking error: one retry on the same connection, then a full restart of the underlying bridge process if still stuck, throttled to 3 restarts per 2 minutes. It's optional (needs Node.js) and the launcher falls straight back to talking to the bridge directly without it. See the README for details.
Run this — do not improvise
Setup and health checks have a deterministic implementation shipped next to this file as
rapp_copilot_in_chrome_agent.py. It is stdlib-only Python with no install step. When asked to
install, verify, or diagnose the bridge, execute it and use its output verbatim — do not guess at
the state of the system:
python3 rapp_copilot_in_chrome_agent.py '{"action": "<action>"}' # doctor | status | install | uninstall
python3 rapp_copilot_in_chrome_agent.py '{"action": "doctor"}' # full check + live Chrome round trip
python3 rapp_copilot_in_chrome_agent.py '{"action": "status"}' # fast check, no browser traffic
python3 rapp_copilot_in_chrome_agent.py '{"action": "install"}' # launcher + MCP registration + skill
python3 rapp_copilot_in_chrome_agent.py --tool # emit the JSON tool contract
doctor verifies all seven links in the chain — Claude binary, native-host manifest, Chrome
extension, launcher, MCP registration, MCP handshake, and a live round trip that reaches a real tab.
Loading the tools (do this first)
The 22 browser tools arrive deferred in Copilot CLI. Load them with a single tool-search call, batching every tool you expect to need — each extra search costs a round trip:
tool_search_tool pattern: "rapp-copilot-in-chrome"
Or target a subset: "tabs_context_mcp|browser_batch|navigate|read_page|computer".
The two rules that matter
- Always call
tabs_context_mcpwithcreateIfEmpty: truefirst. Nearly every other tool requires atabId, and tabs only exist inside this session's tab group. Skip it and you getNo tab available. - Prefer
browser_batchover single calls. It runs a whole sequence in ONE round trip, executing sequentially and stopping on the first error. Batch whenever you can predict two or more steps ahead.
// 1. get a tabId
tabs_context_mcp { "createIfEmpty": true }
// -> {"availableTabs":[{"tabId":1363872857,...}],"tabGroupId":249531617}
// 2. batch the rest
browser_batch {
"actions": [
{ "name": "navigate", "input": { "url": "https://example.com", "tabId": 1363872857 } },
{ "name": "get_page_text", "input": { "tabId": 1363872857 } }
]
}
Tools (22)
Tabs — tabs_context_mcp (list/create group), tabs_create_mcp, tabs_close_mcp
Navigation — navigate (url, or back/forward), resize_window
Reading — get_page_text (article text; best for reading), read_page (accessibility tree with
ref_ids; supports filter: interactive, depth, max_chars), find (natural-language element
lookup returning refs)
Interaction — computer (screenshot, click, type, key, scroll, drag by coordinate),
form_input (set a value by ref), file_upload (by ref — never click a file input, it opens a
native dialog you cannot see), upload_image
Scripting — javascript_tool (REPL semantics in page context: top-level await works and the
last expression is returned — do not write return)
Debugging — read_console_messages (pattern filter), read_network_requests (urlPattern
filter)
Orchestration — browser_batch
Recording — gif_creator (start/stop/export an animated GIF of the session)
Shortcuts — shortcuts_list, shortcuts_execute
Browser selection — list_connected_browsers, select_browser (by deviceId), switch_browser
(broadcast a pairing request, wait for the user to click Connect)
Network tracking is lazy.
read_network_requestsonly begins recording the first time it is called on a tab, so a page that already loaded shows nothing. Call it once before navigating or triggering the requests you want, then call it again to read them. Requests also clear when the tab navigates to a different domain.
Choosing a reading tool
- Reading an article, docs page, or any prose ->
get_page_text - Need to click or fill something ->
find(natural language) orread_pagewithfilter: "interactive"to getrefs - Need pixel coordinates, or the page is canvas/visual ->
computerscreenshot first read_pagetruncates at 50k chars; narrow withref_idordepthrather than raisingmax_charsblindly
Safety
This drives the user's actual browser with their real authenticated sessions. Anything it clicks, submits, purchases, sends, or deletes happens as the user. Confirm before any destructive or irreversible action — sending messages, submitting payments, deleting data, changing account settings. Prefer opening a fresh tab per task rather than reusing tabs the user is actively working in.
Troubleshooting
| Symptom | Fix |
|---|---|
No tab available |
Call tabs_context_mcp { createIfEmpty: true } first, then pass tabId. |
Cannot access a chrome:// URL |
New tabs start on chrome://newtab; navigate somewhere real first. |
read_network_requests empty |
Tracking starts on first call — call it, then navigate/act, then read. |
| Tools not listed | They are deferred — run the tool search above. |
| Server missing | copilot mcp get rapp-copilot-in-chrome should report Status: Enabled. |
Cannot locate claude |
Set RAPP_CHROME_CLAUDE_BIN to its absolute path. |
| Browser not responding | Extension may be disconnected — list_connected_browsers, then select_browser or switch_browser. |
| A call hangs/times out repeatedly | The supervisor already retries and restarts once automatically. If a response mentions the bridge was restarted, call tabs_context_mcp again before anything else — tab/browser selection was reset. |
Run python3 rapp_copilot_in_chrome_agent.py '{"action": "doctor"}' before debugging by hand — it
names the exact broken link in the chain.
<!-- toaster:generated:begin -->
Parameters
The typed contract this capability answers to (JSON Schema — the deterministic layer):
{
"type": "object",
"properties": {
"action": {
"type": "string",
"description": "Derived from `<action>` used in the documented command at line 30."
}
},
"required": []
}
<!-- toaster:generated:end -->
<!-- toaster:generated:begin -->
Deterministic steps
Lifted verbatim from the procedure above by toaster.py toast. Run them in order, substituting the typed parameters; do not paraphrase:
python3 rapp_copilot_in_chrome_agent.py '{"action": "doctor"}'
python3 rapp_copilot_in_chrome_agent.py '{"action": "<action>"}' # doctor | status | install | uninstall
python3 rapp_copilot_in_chrome_agent.py '{"action": "doctor"}' # full check + live Chrome round trip
python3 rapp_copilot_in_chrome_agent.py '{"action": "status"}' # fast check, no browser traffic
python3 rapp_copilot_in_chrome_agent.py '{"action": "install"}' # launcher + MCP registration + skill
python3 rapp_copilot_in_chrome_agent.py --tool # emit the JSON tool contract
<!-- toaster:generated:end -->
<!-- rci-capsule:v1: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 -->
04 同仓库更多 Skills
rappterbook
Use Rappterbook with the AI you already have. Read the public network, choose the authorized GitHub identity, register with verified receipts, read complete conversations, contribute useful evidence, and return to real replies. Use for Rappterbook participation, onboarding, contribution repair, or explicitly approved outreach.
rappterbook
Use Rappterbook with the AI you already have. Read the public network, choose the authorized GitHub identity, register with verified receipts, read complete conversations, contribute useful evidence, and return to real replies. Use for Rappterbook participation, onboarding, contribution repair, or explicitly approved outreach.
writing-brief
Turn a rough idea into a one-page writing brief: audience, promise, outline, tone, and the first sentence. Use when someone has a topic but no structure yet.
rapp-work
Safely verify, inspect, discover, scaffold, update, or migrate RAPP Work through the exact canonical SDK and RAPP/1 checkout. Use when someone wants the same plan-first RAPP Work workflow from Claude Code, GitHub Copilot CLI, or another Agent Skills host without copying protocol logic into the host.
rapp-skills
Turn a Python agent file into a skill anyone can use, turn a skill into one Python file that can run on a server, check a skill for problems, show nothing was lost either way, and run a skill's code locally. Use when a user hands over an agent file or a skill and wants to use it here, share it as one file, or take it somewhere else.
hive-network
Use a public Hive Hub organization seed on your own device, coordinate scoped work with your existing AI, and prepare reviewed contributions for the network. Discover real seed packages, verify their contents, plan native RAPP Work setup, work a ready task, and share only owner-approved public artifacts.