Official description3 sourcesMCP servers3d ago

PUIUX Pilot

Auto-configures Claude Code hooks, MCPs, and skills for any project

View on GitHub ↗Direct link · no redirect page

Official descriptionAuto-configures Claude Code hooks, MCPs, and skills for any project. Scans 95+ project types, selects from 28+ hooks, scores quality (0-100, A-F). npm i -g puiux-pilot

Source: collected from the ccplugins/awesome-claude-code-plugins list · type: MCP server
Harvest history: first seen 2026-10-14 · last seen 2026-10-16 · seen 4× · listed by 3 sources
Score: score 303 · 3 sources · mcp

FAQ (answers come from our records)

How do I install it?

We did not find an install command for this repository (our install coverage is only about 4%). Check its README or releases page.

Who is it for?

We filed it under these tags: MCP 服务. The tags come from the entry itself, not a taxonomy we imposed.

Who maintains it?

first catalogued 2026-10-14；last seen 2026-10-16。There is no official certification or quality rating here — we record sources and dates only.

Third-party quality signals

One command to scan, configure, and optimize any project for Claude Code.

以下摘录自该仓库的 README（片段） · 在 GitHub 读完整 README →

Getting Started | Commands | Safety Model | Contributing
---
Adaptive AI coding assistant configurator for Claude Code.
Scans your project, detects the stack (95+ project types), selects the right hooks/MCPs/skills, and configures everything automatically.
Getting Started
More commands
Commands
| Command | What it does |
|---------|-------------|
| init | Scan project, select hooks/MCPs/skills, show plan. **Dry-run by default.** |
| init --apply | Write changes (backup created first). |
| init --force | Overwrite user-modified hooks. |
| scan [path] | Analyze project without changing anything. Shows Project DNA. |
| score [path] | Quality assessment: 6 dimensions, 0-100, A-F grade. |
| doctor | Health check: broken hooks, missing scripts, timeouts, orphans. |
| hooks list | List all hooks with status. |
| translate | **[EXPERIMENTAL]** Translate config between AI coding tools. |
| eject | Clean removal. Restores original state from backup. |
Safety Model
- **Dry-run by default** — init shows what would change without writing anything.
- **Backup before every apply** — stored in ~/.puiux-pilot/backups/ /.
- **Atomic writes** — temp file + rename; no partial writes on crash.
- **Rollback on failure** — if any step fails, previous state is restored.
- **Manifest-scoped uninstall** — eject only removes hooks Pilot installed. User hooks, permissions, and MCP configs are never touched.
- **No cloud dependency** — everything runs locally.
What Gets Installed
Based on your project type, Pilot selects from:
- **28 hook scripts** across 6 events (PreToolUse, PostToolUse, Stop, SessionStart, PreCompact, UserPromptSubmit)
- **7 profiles**: full-stack, api-only, library, security-first, mobile, startup-speed, infra
- **MCP auto-detection**: ESLint, Prisma, Tailwind CSS, icons8, lottiefiles — based on actual dependencies
- **Skills**: clean-code, e2e-delivery, code-quality, design-standards, and more
Example: a Next.js webapp gets 28 hooks + 4 MCPs. A Node API gets 18 hooks + 0 MCPs.
CI
GitHub Actions workflow runs on every push to main:
| Job | OS | What |
|-----|-----|------|
| build-and-test | ubuntu, windows, macos × node 18, 22 | Build + unit tests + CLI smoke |
| release-gates | ubuntu, macos | Full gates: 72 assertions (cold-start, repo matrix, custom hooks, failure injection) |
| smoke-windows | windows | Lightweight smoke: scan, init, apply, idempotency, doctor, eject |
Run gates locally: bash tools/release-gates.sh (uses sandboxed HOME, never touches real ~/.claude).