ReverseCraft DevTools MCP
A powerful MCP server for browser debugging and reverse engineering, providing AI coding assistants with comprehensive browser automation, JavaScript debugging, and network analysis capabilities.
README
ReverseCraft DevTools MCP
A powerful MCP (Model Context Protocol) server for browser debugging and reverse engineering. Provides AI coding assistants with comprehensive browser automation, JavaScript debugging, and network analysis capabilities.
Features
- JavaScript Debugging: Set breakpoints, step through code, inspect variables, and analyze call graphs
- Network Analysis: Monitor, search, and save network requests with full request/response details
- Script Interception: Replace JavaScript code on-the-fly for testing and reverse engineering
- Page Automation: Navigate pages, interact with elements, and capture screenshots
- Console Monitoring: Access and filter console messages with full stack traces
Installation
Using npx (Recommended)
npx @reverse-craft/rc-devtools-mcp@latest
Using npm
npm install -g @reverse-craft/rc-devtools-mcp
rc-devtools-mcp
MCP Client Configuration
Kiro
Add to your Kiro MCP configuration (.kiro/settings/mcp.json):
{
"mcpServers": {
"rc-devtools": {
"command": "npx",
"args": ["@reverse-craft/rc-devtools-mcp@latest"]
}
}
}
Cursor
Add to your Cursor MCP configuration:
{
"mcpServers": {
"rc-devtools": {
"command": "npx",
"args": ["@reverse-craft/rc-devtools-mcp@latest"]
}
}
}
Claude Desktop
Add to your Claude Desktop configuration (claude_desktop_config.json):
{
"mcpServers": {
"rc-devtools": {
"command": "npx",
"args": ["@reverse-craft/rc-devtools-mcp@latest"]
}
}
}
With Custom Options
{
"mcpServers": {
"rc-devtools": {
"command": "npx",
"args": [
"@reverse-craft/rc-devtools-mcp@latest",
"--headless",
"--viewport", "1920x1080"
]
}
}
}
Quick Start
- Configure your MCP client with the server
- Start a conversation with your AI assistant
- Ask it to navigate to a website and debug JavaScript:
Navigate to https://example.com and set a breakpoint on line 10 of main.js
Or analyze network traffic:
List all network requests and show me the API calls
Available Tools
Input Automation
click - Click on an element
uid(string, required): Element uid from page content snapshotdblClick(boolean, optional): Set to true for double click (default: false)
fill - Type text into input or select option
uid(string, required): Element uid from page content snapshotvalue(string, required): Text to fill or option to select
press_key - Press a key or key combination
key(string, required): Key or combination (e.g., "Enter", "Control+A", "Control+Shift+R")- Supported modifiers: Control, Shift, Alt, Meta
Navigation
navigate_page - Navigate to URL, go back/forward, or reload
type(string, optional): Navigation type:url,back,forward, orreloadurl(string, optional): Target URL (required when type=url)ignoreCache(boolean, optional): Ignore cache on reloadenableDebugger(boolean, optional): Enable JavaScript debugger after navigation (default: true)timeout(number, optional): Navigation timeout in milliseconds
new_page - Create a new browser page
url(string, required): URL to load in the new pageincognito(boolean, optional): Open in incognito windownewWindow(boolean, optional): Open in new windowuserDataDir(string, optional): Independent user data directory pathenableDebugger(boolean, optional): Enable JavaScript debugger (default: true)timeout(number, optional): Navigation timeout in milliseconds
close_page - Close a page by index
pageIdx(number, required): Index of page to close (fromlist_pages)
list_pages - List all open pages
- No parameters
select_page - Select a page as active context
pageIdx(number, required): Index of page to select (fromlist_pages)bringToFront(boolean, optional): Focus and bring page to top
clear_cookies - Clear browser cookies
url(string, optional): Clear cookies for specific domain only (omit to clear all)
Network
list_network_requests - List all network requests
pageSize(number, optional): Maximum requests to returnpageIdx(number, optional): Page number (0-based)resourceTypes(array, optional): Filter by resource types (e.g.,["xhr", "fetch", "script"])includePreservedRequests(boolean, optional): Include requests from last 3 navigations (default: false)
get_network_request - Get details of a specific request
reqid(number, optional): Request ID (omit to use currently selected request in DevTools)
search_network_requests - Search requests by URL, method, status, or content
searchContent(string, optional): Search term in URL, headers, and bodies (returns highlighted snippets)urlPattern(string, optional): URL pattern (substring or regex, e.g.,"api/users"or".*\\.json$")method(string, optional): HTTP method filter (GET,POST,PUT,DELETE,PATCH,HEAD,OPTIONS)statusCode(number, optional): Exact status code filterstatusCodeMin(number, optional): Minimum status code (inclusive)statusCodeMax(number, optional): Maximum status code (inclusive)contentType(string, optional): Response content type filter (substring match, e.g.,"json")resourceTypes(array, optional): Filter by resource typesincludePreservedRequests(boolean, optional): Search in preserved requests (default: false)pageSize(number, optional): Maximum results to returnpageIdx(number, optional): Page number (0-based)
save_network_request - Save request/response to file in raw HTTP format
reqid(number, optional): Request ID (omit to use currently selected request)filePath(string, required): File path to save HTTP transactionresponseBodyFilePath(string, optional): Separate file for binary response bodysaveRequestBody(boolean, optional): Save request body to separate filerequestBodyFilePath(string, optional): File path for request body
save_static_resource - Save static resource to file
reqid(number, required): Request ID containing the resourcefilePath(string, required): File path or directory (extension auto-detected from content type)
Debugging
Breakpoint Management
set_breakpoint - Set a JavaScript breakpoint at a specific line
urlRegex(string, required): Regular expression to match script URL (e.g.,".*main\\.js.*")lineNumber(number, required): Line number (1-based, as shown in editors)columnNumber(number, optional): Target column number (0-based) for smart snappingsnapRange(number, optional): Search range around target column for valid positions (default: 100)condition(string, optional): JavaScript expression - breakpoint only triggers when true
remove_breakpoint - Remove a breakpoint by its ID
breakpointId(string, required): CDP breakpoint ID fromset_breakpointorlist_breakpoints
list_breakpoints - List all active breakpoints
searchTerm(string, optional): Filter breakpoints by ID or URL patternpageSize(number, optional): Maximum breakpoints per pagepageIdx(number, optional): Page number (0-based)
clear_all_breakpoints - Remove all breakpoints on the current page
- No parameters
get_possible_breakpoints - Find valid breakpoint locations (useful for minified code)
urlRegex(string, required): Regular expression to match script URLlineNumber(number, required): Line number to search (1-based)startColumn(number, optional): Start column for search range (0-based)endColumn(number, optional): End column for search range (0-based)maxCount(number, optional): Maximum locations to return (default: 20)
Execution Control
step_into - Step into a function call
maxCallStackDepth(number, optional): Max call stack frames to display (default: 4)contextLines(number, optional): Lines of code context before/after (default: 2)maxLocalVariables(number, optional): Max local variables to display (default: 5)showStatus(boolean, optional): Whether to show debug status (default: true)
step_over - Step over to the next line
maxCallStackDepth(number, optional): Max call stack frames to display (default: 4)contextLines(number, optional): Lines of code context before/after (default: 2)maxLocalVariables(number, optional): Max local variables to display (default: 5)showStatus(boolean, optional): Whether to show debug status (default: true)
step_out - Step out of the current function
maxCallStackDepth(number, optional): Max call stack frames to display (default: 4)contextLines(number, optional): Lines of code context before/after (default: 2)maxLocalVariables(number, optional): Max local variables to display (default: 5)showStatus(boolean, optional): Whether to show debug status (default: true)
resume_execution - Resume paused execution
- No parameters
Inspection & Evaluation
get_debugger_status - Get current debugger state with call stack and variables
frameIndex(number, optional): Call frame index to inspect (default: 0)contextLines(number, optional): Lines of code context (default: 5)maxPropertiesPerScope(number, optional): Max properties per scope to displayskipScopeVariables(boolean, optional): Skip scope variable inspectionuseObjectPreviews(boolean, optional): Use object previews instead of full retrievalmaxOutputLines(number, optional): Maximum output lines (default: 100)maxCallStackFrames(number, optional): Max call stack frames (default: 20)maxLineLength(number, optional): Max characters per line (default: 500)
evaluate_on_call_frame - Evaluate JavaScript expression in a specific call frame
expression(string, required): JavaScript expression to evaluateframeIndex(number, optional): Call frame index (default: 0)maxOutputChars(number, optional): Max output characters (default: 10000)filepath(string, optional): Save result to file instead of displaying
get_scope_variables - Get detailed variable information from a scope
frameIndex(number, optional): Call frame index (default: 0)scopeType(string, optional): Filter by scope type (local,closure,block,script,global, etc.)variableName(string, optional): Get specific variable by namesearchTerm(string, optional): Search variables by namepageSize(number, optional): Max variables per pagepageIdx(number, optional): Page number (0-based)maxDepth(number, optional): Max depth for nested objects (default: 3)maxOutputLines(number, optional): Max output lines (default: 100)saveToFile(string, optional): Save output to file pathmaxLineLength(number, optional): Max characters per line (default: 1000)
save_scope_variables - Save all scope variables to a JSON file
filePath(string, required): Output file pathframeIndex(number, optional): Call frame index (default: 0)includeGlobal(boolean, optional): Include global scope (default: false)
disable_debugger - Disable the debugger and remove all breakpoints
- No parameters
XHR/Fetch Breakpoints
set_xhr_breakpoint - Set a breakpoint on XHR/Fetch requests
urlPattern(string, required): URL substring to match (empty string matches all requests)
remove_xhr_breakpoint - Remove an XHR breakpoint
urlPattern(string, optional): URL pattern to remove (omit to remove all XHR breakpoints)
list_xhr_breakpoints - List all active XHR breakpoints
- No parameters
Script Analysis
save_script_source - Save script source code to file
scriptId(string, optional): Script ID (omit to use currently selected script in DevTools)filePath(string, required): File path to save the source code
analyze_call_graph - Analyze function callers and callees
functionName(string, required): Name of the function to analyzeupstreamDepth(number, optional): Max depth for upstream trace (callers, default: 3, max: 10)downstreamDepth(number, optional): Max depth for downstream trace (callees, default: 3, max: 10)urlPattern(string, optional): Regex pattern to filter scripts by URL
search_functions - Search for functions by name pattern
pattern(string, required): Function name pattern (substring or regex)urlPattern(string, optional): Regex pattern to filter scripts by URLmaxResults(number, optional): Maximum results to return (default: 50)
Script Interception
replace_script - Replace code in scripts matching a URL pattern
urlPattern(string, required): Regex to match script URLs (e.g.,".*main\\.js.*")oldCode(string, required): Original code snippet to replace (must match exactly)newCode(string, required): New code snippet- Note: Changes take effect after page refresh. Rules persist across refreshes.
list_script_replacements - List active script replacement rules
- No parameters
remove_script_replacement - Remove a script replacement rule
ruleId(string, required): Rule ID to remove (fromlist_script_replacements)
clear_script_replacements - Remove all script replacement rules
- No parameters
Console
list_console_messages - List console messages
pageSize(number, optional): Maximum messages to returnpageIdx(number, optional): Page number (0-based)types(array, optional): Filter by message types (e.g.,["log", "error", "warn"])- Available types:
log,debug,info,error,warn,dir,dirxml,table,trace,clear,assert,issue, etc.
- Available types:
includePreservedMessages(boolean, optional): Include messages from last 3 navigations (default: false)savePath(string, optional): Save messages to file instead of displayingmaxLineLength(number, optional): Max characters per line (default: 500)
get_console_message - Get a specific console message by ID
msgid(number, required): Message ID fromlist_console_messages
Screenshots & Snapshots
take_screenshot - Capture a screenshot
format(string, optional): Image format:png,jpeg, orwebp(default:png)quality(number, optional): Compression quality for JPEG/WebP (0-100, higher = better quality)uid(string, optional): Element uid to screenshot (omit for full page)fullPage(boolean, optional): Capture full page instead of viewport (incompatible with uid)filePath(string, optional): Save to file instead of attaching to response
take_snapshot - Take a text snapshot of the page
- No parameters
- Returns accessibility tree representation of the page
Configuration Options
| Option | Description | Default |
|---|---|---|
--cdp-url |
CDP URL to connect to a running Chrome instance | http://127.0.0.1:9222 |
--browser-url |
Connect to a debuggable Chrome instance | - |
--ws-endpoint |
WebSocket endpoint for Chrome connection | - |
--headless |
Run Chrome in headless mode | false |
--executable-path |
Path to custom Chrome executable | - |
--viewport |
Initial viewport size (e.g., 1280x720) |
- |
--user-data-dir |
Chrome user data directory | - |
--channel |
Chrome channel (stable, beta, dev, canary) |
stable |
--isolated |
Use temporary user data directory | false |
--no-sandbox |
Disable Chrome sandboxes (for Docker) | false |
--proxy-server |
Proxy server configuration | - |
--proxy-username |
Proxy authentication username | - |
--proxy-password |
Proxy authentication password | - |
--log-file |
Path to save debug logs | - |
--no-category-emulation |
Disable emulation tools | - |
--no-category-network |
Disable network tools | - |
Examples
Connect to an existing Chrome instance:
npx @reverse-craft/rc-devtools-mcp@latest --browser-url http://127.0.0.1:9222
Run in headless mode:
npx @reverse-craft/rc-devtools-mcp@latest --headless
Use Chrome Canary:
npx @reverse-craft/rc-devtools-mcp@latest --channel canary
Set viewport size:
npx @reverse-craft/rc-devtools-mcp@latest --viewport 1920x1080
Use with proxy:
npx @reverse-craft/rc-devtools-mcp@latest --proxy-server http://proxy.example.com:8080
Requirements
- Node.js 20.19.0+, 22.12.0+, or 23+
- Chrome/Chromium browser (automatically managed or custom installation)
License
Apache-2.0
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。