.cursor — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited .cursor (MCP Server) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
<div align="center">
<img src="packages/extension/icon.png" alt="console-bridge icon" width="128" height="128">
Give your AI assistant access to browser console logs
Capture logs from all tabs, search errors, and equip your AI with project-specific debugging skills.
Quick Start · Features · Tools · Config
</div>
1. 🔌 Passive & Always-On No need to launch a special browser instance or connect to specific tabs. console-bridge captures everything in the background.
2. 🧠 Teachable (Project Skills) Define custom debugging workflows in Markdown. Your AI reads them to understand how to debug your specific application.
3. 🔍 Context Efficient Stop dumping massive log files. Search, filter, and fetch only the relevant lines to save tokens and reduce noise.
| console-bridge | Browser automation tools | |
|---|---|---|
| Capture mode | Passive, always on | On-demand, explicit connect |
| Tab coverage | All tabs simultaneously | Single tab at a time |
| Log storage | 10K logs with TTL | None |
| Search | Regex, keywords, time range | Not supported |
| Skills | Custom debugging skills (.md) | Generic page access |
| Use case | Debugging | UI automation |
Best for: "My app is throwing errors, help me debug" workflows.
console-bridge requires two components: a browser extension (captures logs) and an MCP server (stores/queries logs).
Standard config (works with most MCP clients):
{
"mcpServers": {
"console-bridge": {
"command": "npx",
"args": ["-y", "console-bridge-mcp@latest"]
}
}
}<img src="https://cursor.com/deeplink/mcp-install-dark.svg" alt="Install in Cursor" height="32">
<details> <summary>Claude Code</summary>
claude mcp add console-bridge -s user -- npx -y console-bridge-mcp@latestOr add to ~/.claude.json:
{
"mcpServers": {
"console-bridge": {
"command": "npx",
"args": ["-y", "console-bridge-mcp@latest"]
}
}
}</details>
<details> <summary>VS Code</summary>
Add to your VS Code settings or use the CLI:
code --add-mcp '{"name":"console-bridge","command":"npx","args":["-y","console-bridge-mcp@latest"]}'</details>
<details> <summary>Windsurf</summary>
Follow Windsurf MCP documentation. Use the standard config above.
</details>
<details> <summary>Other clients</summary>
Add the standard config to your MCP configuration file. See MCP documentation for client-specific instructions.
</details>
console-bridge-ext-v*.zip from the latest release~/.console-bridge-extension)chrome://extensionsRestart your MCP client, then try:
Show me recent console errorsIf you see log data, you're all set!
Logs stream continuously from all browser tabs to the MCP server. Filter by level, tab, URL pattern, or time range.
Show error logs from the last 5 minutes
Show logs from localhost:3000
Tail logs from the current tabRegex and keyword search with context lines, AND/OR logic, and exclusions.
Search for "failed" OR "error" excluding "expected"
Search for /api\/users\/\d+/ with 3 lines of contextAI-assisted tab ranking based on your project context — ports, domains, URL patterns.
Suggest which tab is my Next.js appFocus on logs from the current navigation only, ignoring stale logs from before the last page refresh.
Show errors from current session onlyRun JavaScript in page context or query DOM elements directly.
Execute: window.featureFlags.enableDebug = true
Query DOM for '.error-message' elementsClick "Open in Cursor" in the extension popup to instantly open Cursor with a context-aware prompt to analyze the current tab's logs.
Teach your AI assistant how to debug your specific project. Create a .console-bridge/ directory in your project root and add Markdown files with debugging playbooks.
Why use Skills?
Example: Create .console-bridge/auth-debug.md:
---
title: Debug Authentication Flow
description: Steps to diagnose login failures
---
1. Filter logs for "AuthService"
2. Check network requests to /api/login
3. Verify JWT token in local storageUsage:
List available debugging skills
Load the auth debugging skill| Tool | Purpose |
|---|---|
console_tabs | List all tabs or get smart suggestions based on project context |
console_logs | List, get single log, or tail with filtering (level, tab, URL, time) |
console_search | Regex or keyword search with context lines and exclusions |
console_snapshot | Quick summary of recent errors, warnings, and patterns |
console_browser_execute | Run JavaScript in page context |
console_browser_query | Query DOM elements by CSS selector |
console_skills_list | List project-specific debugging playbooks |
console_skills_load | Load a specific debugging skill |
Environment variables:
CONSOLE_MCP_PORT=9847 # WebSocket port
CONSOLE_MCP_DISCOVERY_PORT=9846 # HTTP discovery port
CONSOLE_MCP_MAX_LOGS=10000 # Max logs in memory
CONSOLE_MCP_LOG_TTL_MINUTES=60 # Auto-cleanup (0 = disable)┌─────────────────┐ WebSocket ┌─────────────────┐ stdio ┌─────────────────┐
│ │ (port 9847) │ │ │ │
│ Browser Ext. │ ─────────────────▶ │ MCP Server │ ◀──────────────▶ │ AI Assistant │
│ (all tabs) │ log batches │ (stores/query) │ MCP protocol │ (Cursor/Claude) │
│ │ │ │ │ │
└─────────────────┘ └─────────────────┘ └─────────────────┘
│ │
│ captures console.* │ in-memory storage
│ intercepts errors │ filtering engine
│ sanitizes credentials │ search engine
│ │Data flow:
console.log/warn/error/debugnpm install && npm run build
npm run dev:server # Hot reload server
npm run dev:extension # Hot reload extension
npm test # Run testspackages/
├── server/ # MCP server + WebSocket receiver
├── extension/ # Chrome extension (content script + popup)
└── shared/ # Shared types and Zod schemasMIT
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.