claude-code-starter 0.7.0 → 0.9.0

package/README.md

@@ -7,7 +7,7 @@
 [![Node.js](https://img.shields.io/badge/Node.js-18%2B-green.svg)](https://nodejs.org/)
 [![Bun](https://img.shields.io/badge/Bun-compatible-f472b6.svg)](https://bun.sh/)
 
-Intelligent CLI that bootstraps Claude Code configurations tailored to your project's tech stack.
+Intelligent CLI that uses Claude to deeply analyze your project and generate tailored Claude Code configurations.
 
 ## Quick Start
 
@@ -17,11 +17,28 @@ npx claude-code-starter
 claude
 ```
 
+**Requires [Claude CLI](https://claude.ai/download) to be installed.**
+
 ## What It Does
 
 1. **Analyzes your repository** - Detects languages, frameworks, tools, and patterns
-2. **Generates tailored configurations** - Creates skills, agents, and rules for your stack
-3. **Sets up development workflow** - Task tracking, commands, and methodology guides
+2. **Launches Claude CLI** - Claude reads your actual source files and understands your architecture
+3. **Generates a professional CLAUDE.md** - Project-specific documentation with architecture, conventions, and patterns
+4. **Creates supporting configurations** - Skills, agents, rules, and commands tailored to your stack
+
+## How It Works
+
+```
+npx claude-code-starter
+  |
+  ├── 1. Detect tech stack (languages, frameworks, tools)
+  ├── 2. Generate supporting files (skills, agents, rules, commands)
+  ├── 3. Launch Claude CLI for deep project analysis
+  │     └── Claude reads source files → understands architecture → writes CLAUDE.md
+  └── 4. Done! Run `claude` to start working
+```
+
+The key difference from static scaffolding tools: Claude actually **reads your codebase** and generates documentation specific to your project's architecture, patterns, conventions, and domain knowledge.
 
 ## Tech Stack Detection
 
@@ -29,20 +46,21 @@ Automatically detects and configures for:
 
 | Category       | Detected                                                            |
 | -------------- | ------------------------------------------------------------------- |
-| **Languages**  | TypeScript, JavaScript, Python, Go, Rust, Java, Ruby, and more      |
+| **Languages**  | TypeScript, JavaScript, Python, Go, Rust, Java, Ruby, Swift, Kotlin |
 | **Frameworks** | Next.js, React, Vue, Svelte, FastAPI, Django, NestJS, Express, etc. |
-| **Tools**      | npm, yarn, pnpm, bun, pip, cargo, go modules                        |
+| **Tools**      | npm, yarn, pnpm, bun, pip, cargo, go modules                       |
 | **Testing**    | Jest, Vitest, Pytest, Go test, Rust test                            |
-| **Linting**    | ESLint, Biome, Ruff, Pylint                                         |
+| **Linting**    | ESLint, Biome, Ruff, Pylint                                        |
 
 ## Generated Configurations
 
 Based on your stack, creates:
 
-- **📚 Skills** - Framework-specific patterns (e.g., Next.js App Router, FastAPI endpoints)
-- **🤖 Agents** - Specialized assistants (code reviewer, test writer)
-- **📏 Rules** - Language conventions (TypeScript strict mode, Python PEP 8)
-- **⚡ Commands** - Workflow shortcuts (`/task`, `/status`, `/done`, `/analyze`)
+- **CLAUDE.md** - Comprehensive project documentation generated by Claude's deep analysis of your codebase
+- **Skills** - Framework-specific patterns (e.g., Next.js App Router, FastAPI endpoints)
+- **Agents** - Specialized assistants (code reviewer, test writer)
+- **Rules** - Language conventions (TypeScript strict mode, Python PEP 8)
+- **Commands** - Workflow shortcuts (`/task`, `/status`, `/done`, `/analyze`, `/code-review`)
 
 ## Commands
 
@@ -67,25 +85,41 @@ npx claude-code-starter --help    # Show help
 ## Example Output
 
 ```
-╔═══════════════════════════════════╗
-║   Claude Code Starter             ║
-║   AI-Assisted Development Setup   ║
-╚═══════════════════════════════════╝
+Claude Code Starter v0.4.1
+Intelligent AI-Assisted Development Setup
 
-🔍 Analyzing repository...
+Analyzing repository...
 
-📊 Tech Stack Analysis
+Tech Stack
   Language: TypeScript
   Framework: Next.js
   Package Manager: bun
   Testing: vitest
 
-✅ Configuration complete! (14 files)
+Existing project with 42 source files
+
+Generating supporting configuration...
+
+Created:
+  + .claude/settings.json
+  + .claude/skills/pattern-discovery.md
+  + .claude/skills/nextjs-patterns.md
+  ...
+
+Launching Claude for deep project analysis...
+Claude will read your codebase and generate a comprehensive CLAUDE.md
+
+Claude analysis complete!
+
+Done! (15 files)
 
 Generated for your stack:
-  📚 9 skills (pattern-discovery, iterative-development, security, nextjs-patterns, ...)
-  🤖 2 agents (code-reviewer, test-writer)
-  📏 2 rules
+  CLAUDE.md (deep analysis by Claude)
+  9 skills (pattern-discovery, iterative-development, security, nextjs-patterns, ...)
+  2 agents (code-reviewer, test-writer)
+  2 rules
+
+Next step: Run claude to start working!
 ```
 
 ## Project Structure
@@ -94,7 +128,7 @@ After running, your project will have:
 
 ```
 .claude/
-├── CLAUDE.md           # Project-specific instructions
+├── CLAUDE.md           # Project-specific docs (generated by Claude analysis)
 ├── settings.json       # Permissions configuration
 ├── agents/             # Specialized AI agents
 │   ├── code-reviewer.md
@@ -122,15 +156,13 @@ After running, your project will have:
     └── task.md         # Current task tracking
 ```
 
-## How It Works
-
-### Architecture
+## Architecture
 
 ```mermaid
 flowchart LR
     A[CLI] --> B[Analyzer]
     B --> C[Generator]
-    C --> D[.claude/]
+    B --> D[Claude CLI]
 
     B -->|reads| E[package.json]
     B -->|reads| F[config files]
@@ -140,15 +172,16 @@ flowchart LR
     C -->|creates| I[agents/]
     C -->|creates| J[rules/]
     C -->|creates| K[commands/]
+
+    D -->|reads| L[source files]
+    D -->|creates| M[CLAUDE.md]
 ```
 
 ### Artifact Generation
 
-All artifacts are **generated dynamically** based on your detected tech stack. There are no static templates - content is created specifically for your project.
-
-| Artifact Type | Generation Logic |
-|---------------|------------------|
-| **CLAUDE.md** | Generated with project name, description, and detected stack |
+| Artifact Type | Generation Method |
+|---------------|-------------------|
+| **CLAUDE.md** | Claude CLI deep analysis of your actual source files |
 | **settings.json** | Generated with safe default permissions |
 | **Skills** | Core skills + framework-specific patterns (if detected) |
 | **Agents** | Code reviewer and test writer agents |
@@ -166,20 +199,6 @@ When running on an existing project with `.claude/` configuration:
 | **With `-f` flag** | Overwritten |
 | **state/task.md** | Always preserved |
 
-### Detection Sources
-
-The CLI examines these files to detect your stack:
-
-| Source | Detects |
-|--------|---------|
-| `package.json` | JS/TS frameworks, testing, linting |
-| `pyproject.toml` / `requirements.txt` | Python frameworks, tools |
-| `go.mod` | Go and its web frameworks |
-| `Cargo.toml` | Rust and its frameworks |
-| `*.lockfile` | Package manager (bun.lockb, yarn.lock, etc.) |
-| `.github/workflows/` | CI/CD platform |
-| Config files | Linters, formatters, bundlers |
-
 ### Framework-Specific Patterns
 
 When a framework is detected, additional skills are generated:
@@ -190,96 +209,16 @@ When a framework is detected, additional skills are generated:
 | React | `react-components.md` - Hooks, component patterns |
 | FastAPI | `fastapi-patterns.md` - Async endpoints, Pydantic |
 | NestJS | `nestjs-patterns.md` - Modules, decorators, DI |
+| SwiftUI | `swiftui-patterns.md` - Declarative UI patterns |
+| UIKit | `uikit-patterns.md` - View controller patterns |
+| Vapor | `vapor-patterns.md` - Server-side Swift |
+| Jetpack Compose | `compose-patterns.md` - Compose UI patterns |
+| Android Views | `android-views-patterns.md` - XML views |
 
-### Artifact Examples
-
-<details>
-<summary><b>📚 Skill Example</b> (pattern-discovery.md)</summary>
-
-```markdown
----
-description: "Systematic approach to discovering patterns in unfamiliar codebases"
----
+## Requirements
 
-# Pattern Discovery
-
-## When to Use
-- Starting work on unfamiliar code
-- Investigating "how does X work here?"
-- Before making changes to existing systems
-
-## Approach
-1. Find entry points (main, index, app)
-2. Trace data flow through the system
-3. Identify naming conventions
-4. Map directory structure to responsibilities
-```
-</details>
-
-<details>
-<summary><b>🤖 Agent Example</b> (code-reviewer.md)</summary>
-
-```markdown
----
-name: "Code Reviewer"
-description: "Reviews code for quality, patterns, and potential issues"
-model: "sonnet"
-tools: ["Read", "Glob", "Grep"]
----
-
-# Code Reviewer Agent
-
-You are a code reviewer focused on:
-- Code quality and readability
-- Potential bugs and edge cases
-- Performance considerations
-- Security vulnerabilities
-```
-</details>
-
-<details>
-<summary><b>📏 Rule Example</b> (typescript.md)</summary>
-
-```markdown
----
-paths: ["**/*.ts", "**/*.tsx"]
----
-
-# TypeScript Conventions
-
-- Use strict mode (`"strict": true`)
-- Prefer `interface` over `type` for object shapes
-- Use explicit return types on exported functions
-- Avoid `any`, prefer `unknown` for truly unknown types
-```
-</details>
-
-<details>
-<summary><b>⚡ Command Example</b> (task.md)</summary>
-
-```markdown
----
-description: "Start or switch to a new task"
----
-
-# /task
-
-Update `.claude/state/task.md` with:
-- Task description from user input
-- Status: In Progress
-- Empty decisions/notes sections
-- Clear next steps
-```
-</details>
-
-## Tip
-
-Add `.claude/` to your global gitignore:
-
-```bash
-echo ".claude/" >> ~/.gitignore
-git config --global core.excludesfile ~/.gitignore
-```
+- **Node.js 18+** (for running via npx)
+- **[Claude CLI](https://claude.ai/download)** (for deep project analysis)
 
 ## CI/CD
 
@@ -331,19 +270,9 @@ Configure these in your GitHub repository settings:
 | `GITHUB_TOKEN` | Auto-provided by GitHub | Creating releases |
 
 To create an npm token:
-1. Go to [npmjs.com](https://npmjs.com) → Access Tokens
+1. Go to [npmjs.com](https://npmjs.com) -> Access Tokens
 2. Generate a new "Automation" token
-3. Add it as `NPM_TOKEN` in GitHub repo → Settings → Secrets
-
-### Branch Protection (Recommended)
-
-Configure these settings for the `main` branch:
-
-- ✅ Require a pull request before merging
-- ✅ Require status checks to pass before merging
-  - Required checks: `lint`, `typecheck`, `test`, `build`, `pr-check-complete`
-- ✅ Require branches to be up to date before merging
-- ✅ Do not allow bypassing the above settings
+3. Add it as `NPM_TOKEN` in GitHub repo -> Settings -> Secrets
 
 ## Development
 

package/dist/cli.js

@@ -4335,7 +4335,7 @@ async function main() {
     pc.gray("Your CLAUDE.md was generated by deep analysis - review it with: cat .claude/CLAUDE.md")
   );
 }
-var isMain = process.argv[1] === fileURLToPath(import.meta.url);
+var isMain = fs3.realpathSync(process.argv[1]) === fileURLToPath(import.meta.url);
 if (isMain) {
   main().catch((err) => {
     console.error(pc.red("Error:"), err.message);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "claude-code-starter",
-  "version": "0.7.0",
+  "version": "0.9.0",
   "description": "A lightweight starter kit for AI-assisted development with Claude Code",
   "keywords": [
     "claude",
@@ -29,7 +29,9 @@
     "build": "bun run tsup",
     "dev": "bun run tsup --watch",
     "start": "bun dist/cli.js",
-    "test": "bun test",
+    "test": "bun test src/cli.test.ts",
+    "test:e2e": "bun run build && bun test src/cli.e2e.test.ts",
+    "test:all": "bun run build && bun test",
     "typecheck": "bun run tsc --noEmit",
     "lint": "biome lint src/",
     "lint:fix": "biome lint --write src/",