Scan your code without leaving your editor.
V1be Check plugs into your AI coding assistant over MCP. Ask it to scan the project you’re building and get a launch verdict, a Ship Score, and copy-paste fixes back inline — then let it apply them and re-scan.
How it works
1. Add the MCP server
Drop the V1be Check MCP server into your client config — Cursor, Claude Code, Codex, Cline, Windsurf, or any MCP-capable agent. One block, one API key.
2. Ask it to scan
Tell your assistant “run a V1be Check scan on this project.” It zips the code you’re working on (or a GitHub repo) and sends it to the scanner.
3. Fix inline
Get the verdict, Ship Score, and per-finding fix prompts back in the chat — your assistant applies them and re-scans. No browser, no copy-paste.
Add it in one block
Paste this into your client’s MCP config. After signing up, generate your key at Dashboard → Settings → API keys — it scopes to your workspace only.
{
"mcpServers": {
"vibe-check": {
"command": "npx",
"args": ["-y", "@vibe/mcp"],
"env": {
"VIBE_CHECK_API_KEY": "vck_your_key_here",
"VIBE_CHECK_URL": "https://v1becheck.com"
}
}
}
}A local scan zips your project (skipping node_modules/.git) and uploads it so the scanner can read it — including .env files, since that’s how leaked secrets are caught. The upload is deleted right after the scan and secrets are redacted before any AI step. Prefer a GitHub scan to send only committed files.