Agent Audit MCP Server
Lighthouse-powered MCP server for AI coding agents, Core Web Vitals, technical SEO, accessibility, and GEO/LLM visibility audits. Its 3 documented tools cluster into fast, reliable, analyze. It runs locally over stdio via the published package.
People connecting coding agent tools to Claude, Cursor, VS Code, or another MCP client. The project is written in TypeScript.
VERIFIED ACTIVE
LAST COMMIT 2026-06-28 · ★ 5 · #126 OF 194 MAINTAINED CODING AGENT · VERIFIED 2026-08-25
MIT · TypeScript servers · how we verify → /methodology
01 · Install Agent Audit
before you install - you'll need
Set LIGHTHOUSE_MCP_ALLOW_LOCALHOST, LIGHTHOUSE_CHROME_NO_SANDBOX before connecting.
Claude Code
claude mcp add fullstackdegen-agent-audit -- npx -y @fullstackdegen/agent-audit Claude Desktop / Cursor / VS Code - add to config
{
"mcpServers": {
"fullstackdegen-agent-audit": {
"command": "npx",
"args": [
"-y",
"@fullstackdegen/agent-audit"
]
}
}
} Same JSON for Cursor. For VS Code, rename the top-level key from `mcpServers` to `servers`.
Using another client? Same JSON, different key
Claude Desktop · mcpServers
Cursor · mcpServers
VS Code · servers
Windsurf · mcpServers
Zed · context_servers
Cline · mcpServers
Roo Code · mcpServers
Continue · mcpServers
LibreChat · mcpServers
Gemini CLI · mcpServers
Codex CLI · mcp_servers
Full setup guides: every client.
02 · Evidence
Security posture
What to check before giving this server access to your agent - from the registry, GitHub, and our own probes. We don't score safety; we show what's verifiable.
runs as local process (stdio) - runs on your machine with your user's permissions
license MIT - declared in the repository
npm package @fullstackdegen/agent-audit - published under the repo owner's npm scope (@fullstackdegen)
registry namespace io.github.fullstackdegen is GitHub-verified and matches the repo owner
03 · What Agent Audit can do
Prose above is summarized from the project's README and registry record - no invented capabilities.
What you can build
An agent gets 3 documented tools across fast, reliable, analyze, including fast, reliable, analyze_website_performance.
The 3 tools it gives your agent
Extracted from the project's README - what agent-audit lets an agent do.
- fast
- - one mobile run and one desktop run.
- reliable
- - three runs per profile, medians, and variability ranges. This is
- analyze_website_performance
- - Runs Lighthouse and site intelligence against a target URL:
Latest releases
v0.3.0 · 2026-06-22
Summary · Publish Agent Audit under the npm scope @fullstackdegen/agent-audit because the unscoped agent-audit name is blocked by npm similarity policy. · Keep the CLI binary and MCP server key as agent-audit. · Add…
v0.1.2 · 2026-06-16
Added the shorter local development startup command: npx -y mcp-server-lighthouse --local. · Added --help and unknown-option handling before MCP stdio startup. · Updated MCP server metadata to report version 0.1.2.…
v0.1.1 · 2026-06-16
Added explicit localhost audit mode with LIGHTHOUSE_MCP_ALLOW_LOCALHOST=true for developer machines. · Documented setup for Claude Desktop, Claude Code, Codex, VS Code/GitHub Copilot, Cursor, and generic MCP clients.…
04 · Who maintains Agent Audit
agent-audit is maintained by fullstackdegen. It's the only MCP server we track from this author; the repo dates to Jun 2026.
05 · Facts
- repository
- github.com/fullstackdegen/agent-audit
- category
- coding agent - ranked #126 of 194 actively-maintained coding agent servers as of 2026-08-25.
- release cadence
- 4 releases in the last 90 days (latest 2026-06-22)
- registry
- io.github.fullstackdegen/agent-audit (active, first published 2026-06-28)
- packages
- npm:@fullstackdegen/agent-audit
06 · Agent Audit FAQ
Is Agent Audit still maintained?
Yes - as of 2026-08-25, its last commit was 2026-06-28 and it shipped 4 releases in the last 90 days. We re-verify nightly.
What can Agent Audit do?
An agent gets 3 documented tools across fast, reliable, analyze, including fast, reliable, analyze_website_performance.
How do I install Agent Audit?
Run `npx -y @fullstackdegen/agent-audit`. The README documents 2 environment variables (LIGHTHOUSE_MCP_ALLOW_LOCALHOST, LIGHTHOUSE_CHROME_NO_SANDBOX) to set first. Set LIGHTHOUSE_MCP_ALLOW_LOCALHOST, LIGHTHOUSE_CHROME_NO_SANDBOX before connecting. You can also paste the ready-made client config above.
Does Agent Audit run locally?
Yes - it's a stdio server: it runs on your machine (via npx) with your user's permissions. Your data stays local unless the server itself calls external APIs.
07 · Alternatives to Agent Audit
Alternatives to Agent Audit
Maintained coding agent servers if Agent Audit isn't the fit.
- Serena A powerful toolkit for coding, providing semantic retrieval and editing capabilities. ★ 28,461 · 2026-08-20
- Desktop Commander MCP server for terminal commands, file operations, and process management ★ 9,387 · 2026-08-21
- Ouroboros Pins an acceptance spec; the verify command and expected output never enter the success contract. ★ 5,649 · 2026-08-25
- Socraticode MCP server for local codebase indexing, semantic search, and code dependency graphs. ★ 3,274 · 2026-08-24
- Magic Cloud Self-hosted low-code platform (Docker); its MCP endpoint runs on your own instance, not via npx. ★ 1,180 · 2026-08-24
- LLM Sandbox Securely run LLM-generated code in isolated containers across 7 languages and 3 container backends. ★ 1,111 · 2026-08-23
Pairs well with
Servers that cover what Agent Audit doesn't - only shown when the pairing reason fits the companion.
- Context7 → documentation companion documentation · ★ 61,171
- GitLab MCP → documentation companion documentation · ★ 1,918
- Storybook MCP Addon → developer tools companion developer tools · ★ 268
- Xcode Build MCP → developer tools companion developer tools · ★ 6,274
- Server Filesystem → filesystem companion filesystem · ★ 38,704
- X402 List MCP → filesystem companion filesystem · ★ 58
More coding agent MCP servers · Google Jules MCP Server · Si Git · Gjalla · EchoVault Memory
More TypeScript MCP servers · Fundzwatch · Qubic MCP · Awesome Mineflayer MCP · Gamedev Pl · Microcharts · see all