ccboot-mcp-server
ccboot doesn't just set up Claude Code — it prevents Claude from shipping broken code.
An MCP server that generates, validates, and manages Claude Code configuration artifacts with built-in enforcement. When you run ccboot_init_project, you get a project where:
- Claude cannot mark a task complete without tests passing
- Claude cannot commit code with empty catch blocks or
any types
- Claude cannot skip error handling or leave TODO comments in finished code
- Claude gets blocked by hooks if it tries dangerous operations
- Claude gets audited by a code quality scanner that catches shortcuts
Why v1.2.0?
v1.0 generated files. v1.2 generates files that enforce quality standards.
| Generic CLAUDE.md templates | Opinionated templates with mandatory practices |
| Placeholder hook configs | Working enforcement hooks that block bad code |
| "Add your instructions here" | Specific, actionable instructions that Claude follows |
| Manual code review | Automatic ccboot_audit_code_quality scans |
Features
- Enforcement by Default — Every project gets test-enforcement, linting, and secret-detection hooks
- Code Quality Auditing — Scan for empty catch blocks,
any types, TODO comments, hardcoded secrets, large files, long functions
- Quality Scoring — Get a 0-100 score for your codebase with actionable fix recommendations
- Full Project Bootstrap — Generate complete Claude Code configurations tailored to your tech stack
- Smart Detection — Automatically detects frameworks, build tools, and project structure
- 16 Specialized Tools — Scaffolding, knowledge management, validation, auditing, and enterprise compliance
- 18+ Templates — Battle-tested templates for CLAUDE.md, skills, agents, and hooks
- Compliance Packs — HIPAA, SOC2, PCI-DSS, and SOX compliance configurations
- Security Hooks — Block dangerous commands (rm -rf, force push) and detect secret leaks
Installation
Quick Install (Recommended)
claude mcp add ccboot -- npx @vaspera/ccboot-mcp-server
Manual Installation
npm install -g @vaspera/ccboot-mcp-server
claude mcp add ccboot -- ccboot-mcp-server
Local Development
git clone https://github.com/vaspera/ccboot-mcp-server.git
cd ccboot-mcp-server
npm install
npm run build
claude mcp add --scope local ccboot -- npx tsx src/index.ts
Verification
claude mcp list
/mcp
Tools Reference
Tier 1: Core Scaffolding
ccboot_init_project | Full project bootstrap with CLAUDE.md, skills, hooks, and settings |
ccboot_generate_claudemd | Create or update CLAUDE.md with smart tech stack detection |
ccboot_create_skill | Scaffold a skill with YAML frontmatter and directory structure |
ccboot_create_agent | Create a subagent definition with role-based defaults |
ccboot_create_hook | Generate hook configurations for any lifecycle event |
ccboot_create_command | Create a slash command as a user-invocable skill |
Tier 2: Knowledge & Management
ccboot_create_knowledge | Add reference documents (ADR, runbook, API, schema, style) |
ccboot_create_adr | Generate Architecture Decision Records with auto-numbering |
ccboot_list_artifacts | List all Claude Code artifacts with validation status |
ccboot_validate_config | Deep validation with auto-fix capability |
ccboot_audit_context | Analyze context budget and token consumption |
ccboot_audit_code_quality | NEW Scan codebase for quality issues: empty catch blocks, any types, TODO comments, secrets, large files, long functions. Returns quality score (0-100) |
Tier 3: Enterprise & Compliance
ccboot_apply_compliance | Apply HIPAA, SOC2, PCI-DSS, or SOX compliance packs |
ccboot_create_ci_workflow | Generate GitHub Actions workflow for Claude Code PR review |
ccboot_create_security_hook | Block dangerous commands with configurable strictness |
ccboot_generate_mcp_config | Create .mcp.json with recommended service configurations |
Usage Examples
Bootstrap a New Project
Ask Claude: "Initialize this project with Claude Code configuration"
This runs ccboot_init_project which:
- Detects your tech stack (Next.js, FastAPI, Spring Boot, etc.)
- Generates a tailored CLAUDE.md
- Creates
.claude/ directory with settings.json
- Adds recommended skills and hooks
- Sets up .claudeignore
Generate CLAUDE.md for an Existing Project
Ask Claude: "Generate a CLAUDE.md file for this Next.js project"
The tool analyzes your codebase and creates context-aware instructions including:
- Build and test commands from package.json
- Framework-specific architecture rules
- Detected database and ORM configurations
- CI/CD and deployment patterns
Create a Custom Skill
Ask Claude: "Create a code review skill that runs on PR tasks"
ccboot_create_skill({
name: "code-review",
description: "Comprehensive code review with security and performance checks",
invocation: "auto",
context: "fork",
allowed_tools: ["Read", "Grep", "Glob"]
})
Add Security Hooks
Ask Claude: "Add security hooks to block dangerous commands"
ccboot_create_security_hook({
project_path: ".",
strictness: "strict"
})
This blocks:
rm -rf /, rm -rf ~, rm -rf .
git push --force, git push -f
DROP TABLE, DROP DATABASE, TRUNCATE
git reset --hard, chmod 777
curl | sh, wget | sh
Apply Compliance Pack
Ask Claude: "Apply HIPAA compliance configuration"
ccboot_apply_compliance({
project_path: ".",
standard: "hipaa",
scope: "full"
})
This generates:
- Audit logging hooks
- PHI handling skill
- Data classification CLAUDE.md sections
- Security hooks for sensitive data protection
Audit Code Quality
Ask Claude: "Audit the code quality of this project"
ccboot_audit_code_quality({
project_path: ".",
language: "typescript",
output_format: "json"
})
Scans for:
- Empty catch blocks (swallowed errors)
any type usage (type safety violations)
- TODO/FIXME/HACK comments (incomplete work)
- Hardcoded secrets and API keys
- Files over 400 lines
- Functions over 50 lines
- Missing error handling
- Console.log statements in production code
Returns a quality score (0-100) calculated as:
100 - (15 × critical) - (5 × high) - (2 × medium) - (1 × low)
Validate All Configurations
Ask Claude: "Validate my Claude Code configuration"
ccboot_validate_config({
project_path: ".",
fix_mode: "auto_fix"
})
Checks for:
- Missing name/description in skill frontmatter
- Invalid JSON in hooks and settings
- CLAUDE.md exceeding 500 lines
- Broken file references
- Naming convention violations
Supported Tech Stacks
| Next.js | next.config.js, next.config.ts, next.config.mjs |
| React | src/App.tsx, src/App.jsx |
| Vue / Nuxt | vue.config.js, nuxt.config.ts |
| Angular | angular.json |
| Svelte | svelte.config.js |
| Express | app.js, server.js |
| FastAPI | main.py, app/main.py |
| Django | manage.py, settings.py |
| Flask | app.py, wsgi.py |
| Spring Boot | pom.xml, build.gradle |
| Rails | Gemfile, config/routes.rb |
| Laravel | artisan, composer.json |
| Go | go.mod |
| Rust | Cargo.toml |
| .NET | *.csproj, *.sln |
Hook Events
The server supports all 18 Claude Code lifecycle events:
SessionStart | Session begins or resumes | No |
SessionEnd | Session terminates | No |
InstructionsLoaded | CLAUDE.md loaded | No |
UserPromptSubmit | User submits prompt | Yes |
PreToolUse | Before tool execution | Yes |
PostToolUse | After tool succeeds | Yes |
PostToolUseFailure | After tool fails | Yes |
PermissionRequest | Permission dialog appears | Yes |
SubagentStart | Subagent spawned | No |
SubagentStop | Subagent finished | Yes |
Stop | Claude finishes responding | Yes |
Notification | Claude sends notification | No |
TeammateIdle | Agent teammate going idle | Yes |
TaskCompleted | Task marked complete | Yes |
ConfigChange | Configuration changes | Yes |
PreCompact | Before context compaction | No |
WorktreeCreate | Worktree being created | Yes |
WorktreeRemove | Worktree being removed | No |
Templates
CLAUDE.md Templates
nextjs — Next.js + Prisma + Tailwind
react-spa — React SPA with REST/GraphQL
fastapi — Python FastAPI backend
springboot — Java Spring Boot enterprise
monorepo — Multi-package monorepo
Skill Templates
code-review — Comprehensive code review
test-writer — Test generation
documentation — Documentation generation
security-scan — OWASP security checks
performance-audit — Performance analysis
Agent Templates
explore — Read-only codebase investigation
plan — Architecture planning
test-runner — Test execution and analysis
security-auditor — Vulnerability scanning
Hook Templates
format-on-write — Auto-format with Prettier/Black/gofmt
lint-on-write — Auto-lint with ESLint/Ruff/Clippy
test-enforcement — Run tests after code changes and block completion until green
danger-blocker — Block dangerous commands (rm -rf, force push, DROP TABLE)
secret-detector — Detect and block secret leaks before commits
Knowledge Templates
adr — Architecture Decision Records
runbook — Operational runbooks
api — API documentation
schema — Database schema documentation
style — Code style guides
Configuration
Project Settings (.claude/settings.json)
{
"permissions": {
"allow": [
"Bash(npm run build)",
"Bash(npm run test *)",
"Bash(git status)",
"Bash(git diff *)"
],
"deny": [
"Bash(rm -rf *)",
"Read(./.env)",
"Read(./.env.*)"
]
}
}
MCP Configuration (.mcp.json)
{
"mcpServers": {
"ccboot": {
"command": "npx",
"args": ["@vaspera/ccboot-mcp-server"]
}
}
}
Development
npm run build
npm run lint
npm run test
npm run dev
Architecture
src/
├── index.ts # Entry point, server init, transport
├── types.ts # Shared TypeScript interfaces
├── constants.ts # Enums, defaults, limits
├── tools/
│ ├── scaffolding.ts # Core scaffolding tools
│ ├── knowledge.ts # Knowledge management tools
│ ├── management.ts # Validation and audit tools
│ └── enterprise.ts # Compliance and security tools
├── services/
│ ├── file-ops.ts # Atomic writes, path ops
│ ├── template-engine.ts # Variable substitution
│ ├── validation.ts # Config validators
│ ├── project-analyzer.ts # Stack detection
│ └── claudemd-generator.ts # CLAUDE.md generation
└── schemas/
├── common.ts # Shared Zod schemas
└── scaffolding.ts # Tool input schemas
templates/
├── claudemd/ # CLAUDE.md templates
├── skills/ # Skill templates
├── agents/ # Agent templates
├── hooks/ # Hook templates
└── knowledge/ # Knowledge doc templates
Requirements
- Node.js 18+
- Claude Code CLI
- Claude Pro, Max, or API subscription
License
ISC
Author
Vaspera Capital
Built for Claude Code v1.0+