Claude Cli Advanced Starter Pack

Claude CLI Advanced Starter Pack (CCASP)

╔═══════════════════════════════════════════════════════════════════════════════╗
║                                                                               ║
║   ╔═╗╦  ╔═╗╦ ╦╔╦╗╔═╗  ╔═╗╔╦╗╦  ╦╔═╗╔╗╔╔═╗╔═╗╔╦╗  ╔═╗╔╦╗╔═╗╦═╗╔╦╗╔═╗╦═╗       ║
║   ║  ║  ╠═╣║ ║ ║║║╣   ╠═╣ ║║╚╗╔╝╠═╣║║║║  ║╣  ║║  ╚═╗ ║ ╠═╣╠╦╝ ║ ║╣ ╠╦╝       ║
║   ╚═╝╩═╝╩ ╩╚═╝═╩╝╚═╝  ╩ ╩═╩╝ ╚╝ ╩ ╩╝╚╝╚═╝╚═╝═╩╝  ╚═╝ ╩ ╩ ╩╩╚═ ╩ ╚═╝╩╚═       ║
║                                                                               ║
║                           v2.11.4  •  Production Ready                        ║
║                                                                               ║
╚═══════════════════════════════════════════════════════════════════════════════╝

[](https://www.npmjs.com/package/claude-cli-advanced-starter-pack) [](https://nodejs.org) [](https://opensource.org/licenses/MIT)

---

Your entire Claude Code setup — customized for your stack — deployed in one command.

npx claude-cli-advanced-starter-pack init

[Quick Start](#quick-start) • [What You Get](#what-you-get) • 📚 Wiki • [Roadmap](#roadmap)

---

v2.11.4 Feature Update — Neovim Deploy Panel & Session Management

> Feb 2026 — Enhanced Neovim plugin with deploy configuration panel, session reorder controls, and always-on symlink sync for project templates. Plus per-session slash commands and restored minimized session support.

What's New in v2.11.4

Deploy Configuration Panel

New flyout panel in nvim-ccasp:

• Deploy Config — Configure deployment settings from Neovim sidebar
• Service Management — Railway/Cloudflare settings at a glance
• Quick Deploy — Trigger deployments without leaving the editor

Session Reorder & Management

Footer taskbar enhancements:

• Session Reorder Button — Drag-and-drop session ordering in footer
• Per-Session Slash Commands — Each session gets its own command scope
• Active Session Targeting — Commands route to the focused session
• Restored Minimized Sessions — Splash screen recovery for minimized sessions

Always-On Symlink Sync

Template synchronization improvements:

• Symlink Sync — Project templates stay in sync automatically
• No Manual Refresh — Changes propagate without ccasp init
• Dev Mode Compatible — Works seamlessly with worktree workflows

Previous: Competitor Research & MVP Build Mode (v2.11.1)

• 6-Phase Competitive Intelligence in /project-implementation
• B2C MVP vs Enterprise B2B build tier selection
• Auto-Configure Feature Toggles by market type
• Build Tier Awareness in /phase-dev-plan and /create-roadmap

Upgrade Now

# Update to v2.11.4
npm update -g claude-cli-advanced-starter-pack

# Sync new commands to your project
ccasp init

---

v2.11.1 — Competitor Research & MVP Build Mode (click to expand)

> Feb 2026 — New optional competitive intelligence workflow during project setup. Run 6-phase competitor research, validate compliance, and choose between B2C MVP (6-week, freemium) or Enterprise B2B (10-week, full compliance) build tiers.

• Competitor Research Workflow — 6-phase competitive intelligence in /project-implementation
• Build Tier Selection — B2C MVP (6 weeks, freemium) vs Enterprise B2B (10 weeks, full compliance)
• Auto-Configure Feature Toggles — Tier selection auto-stubs/enables multi-tenancy, RBAC, billing
• Neovim Panel — New "MARKET RESEARCH" section with competitor_research toggle
• Build Tier Awareness — /phase-dev-plan and /create-roadmap read buildTier from tech-stack.json

v2.8.1 — Getting Started Onboarding + Help System (nvim-ccasp) (click to expand)

> Feb 2026 — Interactive 8-page Getting Started guide with first-launch detection, ASCII diagrams, try-it actions, and a persistent Help system with 11 searchable wiki topics embedded offline.

• Getting Started Guide — 8-page onboarding with ASCII diagrams, keybinding tables, try-it actions
• First-Launch Detection — Auto-opens on first launch, remembers completion state
• Help System — 11 wiki topics with Telescope fuzzy search + vim.ui.select fallback
• Navigation — Page dots, n/p keys, 1-8 jump, h/l arrow support
• New Commands — :CcaspWelcome, :CcaspHelp, c?, cw
• Sidebar Integration — Help tab now shows Getting Started and Help links

v2.7.0 — Vision Multi-Instance & Decision Engine (click to expand)

> Feb 2026 — Vision Mode now supports multiple concurrent visions with a centralized registry, cross-process file locking, slug uniqueness enforcement, and an intelligent decision engine.

• Vision Registry: Centralized VISION_REGISTRY.json with atomic writes, auto-rebuild
• Decision Engine: Weighted scoring selects plan type (task-list / phase-dev-plan / roadmap / epic / vision-full)
• Slug Uniqueness: generateUniqueSlug() appends -2/-3 on collision
• New Commands: ccasp vision list, ccasp vision cleanup, /vision-list

v2.6.1 — Smart Todo List with AI Intent Routing (click to expand)

> Feb 2026 — New Smart Todo system: quick-add with auto-paraphrasing, interactive menu, AI-powered "Work on item" that suggests the best CCASP workflow (Vision/Epic/Roadmap/Phase-Dev/Task List), plus full Neovim integration with sidebar, floating panel, and Telescope picker.

• /todo-add — Quick-add with auto-paraphrasing to 5-6 word titles
• /todo — Interactive menu: list, navigate, work-on, complete, delete
• AI Intent Routing — Analyzes todo and suggests Vision/Epic/Roadmap/Phase-Dev/Task List/Just do it
• Neovim Integration — Sidebar section, floating panel, Telescope picker

---

🚀 v2.6.0 Feature Update — Smart Site Intel Dev-Scan System

> Feb 2026 — New developer-focused application scanner: per-route health monitoring, data-testid coverage tracking, git-diff incremental scanning, regression detection, and interactive dashboard with 12 MCP tools.

🆕 What's New in v2.6.0

🔬 Smart Site Intel Dev-Scan

Developer-focused per-route application scanning:

• Route Catalog — AST-based route discovery with reverse file→route mapping
• Config Reader — Reads tech-stack.json, .env, playwright.config automatically
• Playwright Scanner — Auth, live DOM testid check, axe a11y, optional Lighthouse
• Git Diff Integration — Incremental scanning with 30% threshold for full scan
• State Persistence — Per-route scores, history (20 entries FIFO), diff tracking
• Quick Check — Static testid coverage in

📊 Dev Scan Dashboard

Interactive web dashboard at http://localhost:3847:

• Health Metrics — Overall score, total routes, testid coverage, last scan
• Sortable Route Table — Sort by path, testid, a11y, perf, health grade
• Color-Coded Grades — A (green) through F (red) per-route health
• Diff Display — Improvements (green) and regressions (red)
• Trend Sparklines — Inline SVG health history per route
• Filter Controls — All routes, failing (

🔌 3 New MCP Tools (12 Total)

Expanded site-intel MCP server:

• site_intel_dev_scan — Full/incremental per-route scan with diffs
• site_intel_quick_check — Fast static testid coverage (no Playwright)
• site_intel_dev_state — Read current state without scanning
• Updated /site-intel slash command with dev-scan actions
• New /site-intel-dashboard dedicated slash command

📦 Dev-Scan Module (8 Files, ~78KB)

Complete scanning subsystem:

• config-reader.js — Project config loading (226 lines)
• route-catalog.js — Route discovery + reverse map (246 lines)
• state-manager.js — Persistent state + history (305 lines)
• git-diff.js — Changed file detection + route mapping (184 lines)
• testid-checker.js — Static analysis, 11 patterns (326 lines)
• scanner.js — Playwright-based scanning (361 lines)
• diff-reporter.js — Categorized diff output (201 lines)

Upgrade Now

# Update to latest
npm update -g claude-cli-advanced-starter-pack

# Sync all new features to your project
ccasp init

---

The Problem

Every new project means:

• ❌ Manually creating .claude/ folders and configurations
• ❌ Copy-pasting commands, hooks, and agents from old projects
• ❌ Forgetting critical settings and enforcement hooks
• ❌ Agents that don't understand your tech stack
• ❌ Hours of setup before you can actually build

The Solution

One command. Fully configured. Stack-aware.

npx claude-cli-advanced-starter-pack init

CCASP scans your project, detects your stack (React? FastAPI? Prisma? Playwright?), and deploys a complete .claude/ folder with:

• ✅ 64 slash commands — deployment, testing, refactoring, GitHub sync, UI testing
• ✅ 44 enforcement hooks — validation, progress tracking, agent orchestration
• ✅ Stack-specific agents — specialists that know your frameworks
• ✅ Smart updates — add features without losing customizations

---

What You Get

| Component | Count | Examples | |-----------|-------|----------| | Slash Commands | 66 | /deploy-full, /pr-merge, /ralph, /ui-test, /research-competitor | | Enforcement Hooks | 44 | Token tracking, agent orchestration, refactor safety, UI test auto-run | | Tech Stack Detection | 55+ | React, Vue, FastAPI, Express, Prisma, Playwright | | Agent Templates | L1/L2/L3 | Orchestrators → Specialists → Workers + UI testing agents | | MCP Integrations | 50+ | Railway, Cloudflare, GitHub, Playwright | | Skills (RAG) | 6 | agent-creator, hook-creator, refactor patterns, nvim-ui-test |

---

AI Constitution Framework

Enforce project-specific coding standards and architecture patterns with automatic validation.

Quick Start

ccasp constitution-init --preset senior    # 5% sampling, all sections
ccasp constitution-init --preset minimal   # 2% sampling, security only
ccasp constitution-init --preset strict    # 15% sampling, maximum coverage

Features

• YAML-based rules at .claude/config/constitution.yaml
• Hook enforcement with configurable sampling (1-in-20 default)
• Sensitive pattern bypass - security rules always checked
• Stop → Explain → Suggest → Refactor violation flow
• Neovim integration - edit constitution from sidebar

Rule Sections

| Section | Examples | |---------|----------| | code_style | Naming, types, imports | | architecture | Patterns, boundaries | | security | Credentials, validation | | performance | Async, memoization | | git | Commits, branches | | dependencies | Versions, security |

See /ai-constitution-framework in Claude Code for full documentation.

---

Landing Page Generator

Automated marketing asset pipeline: capture screenshots, apply device frames, generate GIFs, and build landing page components.

Commands

| Command | Description | |---------|-------------| | /landing-page-generator | Interactive menu: configure, capture, frame, GIF, preview, build, deploy | | /landing-page-generator-executor | Automated phase executor for pipeline tasks | | /landing-page-toggle | Enable/disable landing page in production |

Pipeline

1. Capture - Playwright screenshots of app routes at configurable viewports
2. Frame - Composite into device bezels (MacBook Pro, iPhone 15, iPad, etc.)
3. Generate GIFs - Stitch framed screenshots into animated marketing GIFs
4. Build - Generate React components with the produced assets

Skill: Screenshot Pipeline

The screenshot-pipeline skill package includes reference documentation for:

• Device frames - 12+ device templates (laptop, phone, tablet)
• Viewport presets - Standard configurations for phone/tablet/desktop
• Pipeline config - Full configuration schema with examples

Requires: sharp, gifencoder, canvas NPM packages + playwright-ext MCP server.

---

See It In Action

📸 Screenshots Gallery (click to expand/collapse)

Desktop UI

/menu — Interactive Navigation

Full-featured menu with all options

Project Configuration Settings

GitHub, deployment, tunnels, tokens

Mobile UI (Happy.engineering)

/menu — Mobile Navigation

Single-key navigation, 40-char width

Setup Wizard — Feature Selection

Choose commands, agents, hooks, skills

GitHub Project Setup

One-click repo and Project Board sync

/claude-audit — Analysis Report

CLAUDE.md analysis with score

Audit Score Breakdown

Detailed scoring and summary

Actionable Recommendations

Priority-ordered improvements

---

Highlight Features

🎯 Stack-Aware Agents

Auto-generates agents that know your tech:

• React + Zustand + Tailwind specialist
• FastAPI + SQLAlchemy backend agent
• Playwright E2E testing expert
• PostgreSQL + Prisma database agent

🔄 Ralph Loop Testing

Continuous test-fix cycle until green:

/ralph --watch

Runs tests → Parses failures → Fixes code → Repeats.

Smart Recovery: Every 3rd failed attempt deploys a web search agent to find best-practice solutions for stubborn bugs or implementation issues.

Max 10 iterations. Stops on 3x same failure.

🗺️ Multi-Phase Roadmaps

Coordinate full-stack, multi-feature development:

/create-roadmap "Q1 Feature Sprint"

• Break epics into phased milestones
• GitHub issue tracking per phase
• Customizable Project Board sync
• Dependency graphs with auto-ordering
• Progress dashboard across all phases

🚀 One-Command Deployment

Full-stack deploy to Railway + Cloudflare:

/deploy-full

• Backend and frontend in parallel
• Pre-flight validation
• Automatic rollback on failure

📋 Phased Development

Break complex features into phases:

/phase-dev-plan "Add user authentication"

• 95%+ success probability planning
• PROGRESS.json state tracking
• GitHub issue sync
• L1/L2/L3 agent orchestration

🔀 Safe PR Merges

9-phase merge workflow with safety:

/pr-merge

• Auto-stash uncommitted changes
• Resolve conflicts, CI failures, reviews
• Automatic rollback on any failure

🎫 GitHub Epic System

Multi-issue epic workflows:

/create-github-epic "Auth System"

• Creates parent epic with child issues
• Tracks completion across all issues
• Auto-syncs with Project Board
• Progress comments on epic issue

📱 Mobile-First UI

Works with Happy.engineering:

• 40-character max width
• Card-based layouts
• Single-character inputs
• Auto-detected via environment

👁️ Vision Mode — Autonomous MVP Development

Transform natural language into complete, working MVPs:

ccasp vision init "Build a kanban board with real-time collaboration"

8-Stage Orchestrated Workflow:

1. Initialize — Parse prompt, detect intent, extract features
2. Analyze — Web search for inspiration, discover npm/pip packages, match MCP servers
3. Architect — Generate Mermaid diagrams, API contracts, ASCII wireframes
4. Security — Scan dependencies with npm audit, pip-audit, OSV Scanner
5. Create Agents — Spawn specialized agents for detected tech stack
6. Execute — Autonomous development loop with self-healing
7. Validate — Run tests, verify MVP completeness
8. Complete — Final verification and checkpoint creation

Features:

• 🎯 Natural Language Input — "Build a todo app with React and FastAPI"
• 🔍 Analysis Engine — Similar apps search, tool discovery, MCP matching
• 🏗️ Architecture Planning — Mermaid diagrams, API contracts, state design
• 👁️ Drift Detection — Hook-based observer with automatic plan adjustment
• 🔒 Security Scanning — Pre-install vulnerability detection
• 🤖 Dynamic Agents — Creates specialists based on tech stack
• 🔄 Self-Healing — Automatic test failure resolution
• 📊 Web Dashboard — Real-time status at http://localhost:3847

`

…

Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: evan043
• Source: evan043/claude-cli-advanced-starter-pack
• License: MIT

Install and usage instructions live in the source repository linked above.