Agent Audit v0.3.0

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 schema v1.2 Agent Fix Packs for coding-agent implementation handoff.
• Refresh README, AGENTS.md, CLAUDE.md, workflow docs, examples, and release validation.

Install

npx -y @fullstackdegen/agent-audit

Verification

• npm test: 18 files, 112 tests passed
• npm run check
• npm run build
• npm run validate:release
• npm pack --dry-run --cache /private/tmp/agent-audit-npm-cache

v0.1.2

What's changed

• 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.
• Kept LIGHTHOUSE_MCP_ALLOW_LOCALHOST=true as a fallback for clients that cannot pass CLI arguments.

Verification

• CI passed on Node 20 and Node 22.
• Published npm package: mcp-server-lighthouse@0.1.2.

v0.1.1

What's changed

• 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.
• Removed internal planning files from the public repository tree.
• Pinned Lighthouse to 12.6.1 to preserve Node 20 support with a clean production npm audit.

Verification

• CI passed on Node 20 and Node 22.
• Published npm package: mcp-server-lighthouse@0.1.1.

Lighthouse MCP v0.1.0

Turn Lighthouse audits into coding-agent tasks

Lighthouse MCP is an npm-distributed MCP server that runs repeatable mobile and desktop Lighthouse audits and returns an implementation-ready report for coding agents.

Install

npx -y mcp-server-lighthouse

Highlights

• Mobile and desktop Performance, Accessibility, Best Practices, and SEO audits
• Fast mode with one run per profile and reliable mode with three-run medians
• Structured JSON plus equivalent Markdown
• At most 10 prioritized coding tasks
• Resource, DOM, and console evidence when Lighthouse provides it
• Deterministic suggested actions and measurable acceptance criteria
• Public-network URL policy to reduce SSRF risk

Examples

• JSON report: https://github.com/fullstackdegen/lighthouse-mcp/blob/main/examples/commalabs-fast-report.json
• Markdown report: https://github.com/fullstackdegen/lighthouse-mcp/blob/main/examples/commalabs-fast-report.md

Lighthouse results vary with browser version, hardware, network conditions, and page changes. The included CommaLabs report demonstrates the output shape, not a permanent score claim.

Requirements

• Node.js 20 or later
• Google Chrome or Chromium