MCP Vibe Coding Tools

Complete Autonomous Development Organization - Transform any prompt into a fully functional, production-ready, revenue-generating product with zero human intervention.

🎯 Mission: From Prompt to Profit

This MCP server operates as a complete software development company in a box, providing 100+ tools that enable AI to function as:

• 📋 Requirements Team - Requirements analysis, user stories, acceptance criteria

• 📊 Product Team - Roadmaps, competitive analysis, market research

• 🔬 R&D Team - Technology research, architecture design, proof of concepts

• 🏗️ IT Department - Infrastructure, deployment, monitoring, automation

• 🔒 Security Team - Vulnerability scanning, auditing, compliance, hardening

• 💻 Development Team - Implementation, testing, code review, optimization

• 📚 Documentation Team - Comprehensive docs, API specs, guides

The Revolution

Traditional AI Coding: Generate code → Ask human for validation → Wait for approval → Repeat

MCP Vibe Coding: Prompt → Requirements → Architecture → Implementation → Testing → Security → Deployment → DONE

Philosophy

• ✅ Act, Don't Ask - AI makes decisions and fixes issues automatically

• ✅ Test Everything - Comprehensive validation before shipping

• ✅ Document Relentlessly - Auto-generated, always up-to-date docs

• ✅ Automate Ruthlessly - Scripts and CI/CD for everything

• ✅ Quality First - Production-ready code only

Related MCP server: MCP AI POC

🚀 What Makes This Different

Traditional AI Coding Tools

• Ask for permission before actions

• Require manual testing and validation

• Generate documentation as an afterthought

• Need human intervention for bug fixes

• Create technical debt

MCP Vibe Coding Tools

• Autonomous execution - Fix issues without asking

• Built-in validation - Auto-test, auto-lint, auto-build

• Self-documenting - Generate docs as code evolves

• Iterative fixing - Debug and resolve automatically

• Production quality - Ship only fully validated code

⚠️ Critical: SDK Version Compatibility

ALWAYS ensure package.json matches the installed MCP SDK version!

This project uses @modelcontextprotocol/sdk@^1.24.3 with the modern McpServer API.

If you see v3Schema.safeParseAsync is not a function:

This is usually a client cache issue, not a server issue. See TROUBLESHOOTING.md for full fix.

Quick fix:

1. Clean rebuild: rm -rf dist node_modules && npm install && npm run build

2. Completely quit and restart your MCP client (Claude Desktop, Cursor, etc.)

3. Clear client cache if needed (see troubleshooting guide)

Version mismatches between package.json and node_modules will break the server.

📦 Three MCP Servers

This package provides three separate MCP servers:

1. mcp-vibe-coding-tools (Main Server)

Local development tools for filesystem, git, testing, automation, diagnostics, kubernetes, and RAG.

Entry point: dist/mcp-vibe-coding-tools.js

2. mcp-gitops-tools (GitOps Server)

Remote GitOps operations for GitHub Actions, GitLab CI/CD, and GitLab API.

Entry point: dist/gitops-server.js

3. mcp-gcloud-tools (Google Cloud Server) 🆕

Google Cloud Platform integration with 35+ tools for GKE, BigQuery, Cloud Run, Dataflow, Logging, and more. Uses Application Default Credentials (ADC) for seamless authentication.

Entry point: dist/gcloud-server.js

Documentation:

• GCP_README.md - Complete overview

• GCP_QUICKSTART.md - Quick start guide

• GCP_TOOLS.md - All 35+ tools reference

• GCP_EXAMPLES.md - Real-world examples

📦 150+ Production-Ready Tools Across 20+ Categories

Filesystem Operations (6 tools)

• read_file - Read files with encoding support

• write_file - Create/update files safely

• list_directory - Browse directory trees

• search_files - Find files by glob pattern

• file_info - Get file metadata

• create_directory - Create directory structures

CLI Execution (3 tools)

• execute_command - Run shell commands

• get_environment - Access environment variables

• which_command - Find command locations

Git Operations (9 tools)

• git_status, git_log, git_diff - Repository inspection

• git_branch, git_commit - Version control

• git_push, git_pull, git_clone - Remote operations

• git_stash - Temporary storage

Web & HTTP (4 tools)

• fetch_webpage - Download web content

• parse_html - Extract structured data

• extract_links - Get all links

• download_file - Save remote files

Node.js/npm (5 tools)

• npm_install - Manage packages

• npm_run_script - Execute scripts

• npm_outdated - Check dependencies

• npm_init - Create projects

• read_package_json - Read metadata

Python/pip (5 tools)

• python_create_venv - Virtual environments

• pip_install - Install packages

• pip_freeze - Generate requirements

• python_run_script - Execute code

• python_version - Check version

Testing & Building (4 tools)

• run_tests - Execute test suites

• build_project - Compile/build

• start_dev_server - Run dev servers

• lint_code - Code quality checks

🤖 Automation & Orchestration (5 tools)

validate_project

Run complete project validation suite:

• Linting with auto-fix

• Type checking

• Test execution with coverage

• Build verification

• Auto-iterates until all checks pass

create_validation_script

Generate comprehensive validation scripts:

• Add lint, test, build commands to package.json

• Set up pre-commit hooks

• Configure coverage thresholds

• Create validation pipeline

setup_project_automation

Complete automation setup:

• GitHub Actions / GitLab CI workflows

• Dependabot configuration

• Pre-commit hooks

• Makefile for cross-platform commands

generate_project_docs

Auto-generate documentation:

• CONTRIBUTING.md with dev workflow

• ARCHITECTURE.md with system design

• CHANGELOG.md with version history

• API documentation

fix_common_issues

Detect and auto-fix problems:

• Missing package.json scripts

• Missing .gitignore

• Missing README

• Outdated dependencies

• Fixes issues without asking

🔍 Diagnostics & Logging (8 tools)

get_vscode_problems

Get real-time compilation and linting errors:

• Runs TypeScript compiler to find type errors

• Executes ESLint to detect code quality issues

• Returns structured problem list with file, line, severity

• Filter by file path or severity level

• Essential for autonomous error fixing

read_log_file

Read and parse log files with smart filtering:

• Tail last N lines (like tail -n)

• Filter by log level (ERROR, WARN, INFO, DEBUG)

• Apply custom regex patterns

• Parse structured logs (JSON, timestamp formats)

• Extract timestamps, levels, messages

tail_log_file

Monitor log files for recent activity:

• Get snapshot of recent log entries

• Useful for monitoring build/test output

• View last N lines of any log file

search_logs

Search all logs for specific patterns:

• Recursive search through log directories

• Context lines before/after matches

• Regex pattern support

• Group matches by file

• Find error patterns across entire workspace

find_log_files

Discover all log files in workspace:

• Glob pattern matching (*.log, **/*.log)

• File size and modification time

• Sort by most recent

• Exclude node_modules by default

analyze_error_logs

Deep analysis of log files for errors:

• Extract errors, warnings, exceptions

• Parse stack traces automatically

• Categorize error types

• Count error patterns

• Identify root causes autonomously

watch_log_changes

Monitor log file changes incrementally:

• Read only new content since last position

• Efficient incremental log monitoring

• Get byte position for next read

• Perfect for long-running processes

get_terminal_history

Access recent terminal commands:

• Read zsh/bash history files

• Parse timestamps (zsh extended format)

• Filter by command pattern

• Learn from previous command executions

aggregate_logs

Combine and analyze multiple log files:

• Group by level, file, hour, or day

• Extract time ranges

• Count entries by category

• See big picture across all logs

get_xcode_logs

Get Xcode build logs, crash logs, and simulator logs:

• Build logs from DerivedData

• Simulator runtime logs

• Crash reports from DiagnosticReports

• Device logs (requires libimobiledevice)

get_android_logs

Get Android logcat and build logs:

• Logcat with priority/tag filtering

• Gradle build logs

• Parse structured log format

get_react_native_logs

Get React Native Metro bundler and app logs:

• Metro bundler output

• iOS simulator logs

• Android logcat for RN apps

Kubernetes Operations (7 tools)

kubectl_get_pods

List pods in a namespace with status information.

kubectl_describe_pod

Get detailed information about a specific pod.

kubectl_get_logs

Fetch logs from a pod with filtering options.

kubectl_get_deployments

List all deployments with replica status.

kubectl_get_services

List all services with IP and port information.

kubectl_get_events

Get cluster events for debugging issues.

kubectl_get_resource_status

Get status of any Kubernetes resource type.

Note: All kubectl tools are read-only - for validation and debugging only, no destructive operations.

GitHub Actions (6 tools)

github_list_workflow_runs

List workflow runs with optional filters (status, branch).

github_get_workflow_run

Get details of a specific workflow run.

github_list_workflow_jobs

List all jobs in a workflow run.

github_get_job_logs

Fetch logs for a specific job.

github_list_workflows

List all workflows in a repository.

github_get_workflow_run_logs

Download complete workflow run logs as base64-encoded zip.

Environment Variable Required:

• GITHUB_API_KEY - GitHub Personal Access Token with repo and actions:read scopes

Error Handling: Tools only error when invoked without GITHUB_API_KEY set - they do NOT error on server startup.

RAG (Retrieval Augmented Generation) (6 tools)

Enable local document search and retrieval for AI coding assistants. Perfect for searching through documentation, codebases, and reference materials.

Environment Variables:

• RAG_DOCS_PATH - Required - Path to directory containing documents to index

Note: The index is automatically stored in .rag-index folder inside RAG_DOCS_PATH. Chunk size (1000) and overlap (200) are sensible defaults.

rag_index_documents

Index local documents for semantic search:

• Scans directory for supported file types

• Chunks documents with configurable overlap

• Builds TF-IDF index for semantic similarity

• Incremental indexing (only re-indexes changed files)

• Supports 30+ file extensions (md, ts, js, py, json, yaml, etc.)

rag_search

Search indexed documents using semantic similarity:

• TF-IDF based semantic search

• Returns top-K most relevant chunks

• Configurable minimum similarity threshold

• File path filtering with regex

• Perfect for finding relevant docs and code examples

rag_get_context

Expand context around search results:

• Get more lines before/after a match

• Useful for understanding surrounding code

• Configurable expansion range

rag_list_indexed

List all indexed documents:

• Shows file paths, types, and chunk counts

• Optional detailed chunk information

• Index statistics and metadata

rag_clear_index

Clear the RAG index:

• Requires confirmation flag

• Removes all indexed documents

rag_status

Check RAG system status:

• Shows if RAG is enabled

• Configuration values

• Index statistics if available

Example Configuration:

{
  "mcpServers": {
    "vibe-coding-tools": {
      "command": "node",
      "args": ["/path/to/mcp-vibe-coding-tools/dist/mcp-vibe-coding-tools.js"],
      "env": {
        "WORKSPACE_PATH": "/path/to/project",
        "RAG_DOCS_PATH": "/path/to/docs-to-search"
      }
    }
  }
}

GitLab CI/CD (7 tools)

gitlab_list_pipelines

List pipelines for a project with optional filters.

gitlab_get_pipeline

Get detailed information about a specific pipeline.

gitlab_list_pipeline_jobs

List all jobs in a pipeline.

gitlab_get_job

Get details of a specific job including artifacts info.

gitlab_get_job_trace

Fetch job logs/trace output.

gitlab_list_project_jobs

List all jobs in a project with filtering.

gitlab_get_pipeline_variables

Get variables used in a pipeline execution.

GitLab API (23 tools)

Comprehensive GitLab API integration for groups, projects, issues, and merge requests.

Groups

• gitlab_get_group - Get group details

• gitlab_list_group_subgroups - List subgroups

• gitlab_list_group_projects - List projects in group

• gitlab_list_descendant_groups - List all nested subgroups

Projects

• gitlab_get_project - Get project details

• gitlab_list_projects - List accessible projects

Issues

• gitlab_list_issues - List issues with filters

• gitlab_get_issue - Get issue details

• gitlab_create_issue - Create new issue

• gitlab_update_issue - Update existing issue

• gitlab_list_issue_notes - List issue comments

• gitlab_create_issue_note - Add issue comment

Merge Requests

• gitlab_list_merge_requests - List MRs with filters

• gitlab_get_merge_request - Get MR details

• gitlab_merge_merge_request - Merge an MR

• gitlab_list_mr_changes - Get MR diff

• gitlab_list_mr_notes - List MR comments

• gitlab_create_mr_note - Add MR comment

• gitlab_approve_merge_request - Approve MR

Global Search

• gitlab_search_issues_global - Search issues across all projects

• gitlab_search_merge_requests_global - Search MRs across all projects

• gitlab_search_group_issues - Search issues in a group

• gitlab_search_group_merge_requests - Search MRs in a group

Environment Variables Required:

• GITLAB_API_KEY - GitLab Personal Access Token or Project Access Token with read_api scope

• GITLAB_HOST (optional) - GitLab instance URL (default: https://gitlab.com for GitLab.com, or set to your self-hosted instance)

Error Handling: Tools only error when invoked without GITLAB_API_KEY set - they do NOT error on server startup.

Google Cloud Platform (35+ tools) 🆕

Comprehensive GCP integration using Application Default Credentials (ADC). No API keys required - uses your gcloud authentication.

Authentication & Configuration (7 tools)

• gcloud_auth_login - Authenticate with ADC

• gcloud_auth_list - List authenticated accounts

• gcloud_auth_print_access_token - Generate access token for API calls

• gcloud_auth_print_identity_token - Generate identity token (JWT) for Cloud Run

• gcloud_config_set - Set configuration properties (project, region, zone)

• gcloud_config_get - Get configuration values

• gcloud_config_list - List all configuration

Cloud Logging (2 tools)

• gcloud_logging_read - Read logs with advanced filtering (time, severity, resource)

• gcloud_logging_write - Write log entries

GKE - Google Kubernetes Engine (4 tools)

• gcloud_container_clusters_list - List GKE clusters

• gcloud_container_clusters_describe - Get cluster details

• gcloud_container_clusters_get_credentials - Configure kubectl credentials

• gcloud_container_node_pools_list - List node pools

BigQuery (4 tools)

• gcloud_bq_query - Execute SQL queries with dry-run support

• gcloud_bq_ls - List datasets and tables

• gcloud_bq_show - Show dataset/table details and schema

• gcloud_bq_mk - Create datasets and tables

Dataflow (3 tools)

• gcloud_dataflow_jobs_list - List Dataflow jobs

• gcloud_dataflow_jobs_describe - Get job details and metrics

• gcloud_dataflow_jobs_cancel - Cancel running jobs

Resource Manager (4 tools)

• gcloud_projects_list - List all accessible projects

• gcloud_projects_describe - Get project details

• gcloud_services_list - List enabled/available APIs

• gcloud_services_enable - Enable Google Cloud APIs

Compute Engine (2 tools)

• gcloud_compute_instances_list - List VM instances

• gcloud_compute_instances_describe - Get instance details

Cloud Run (2 tools)

• gcloud_run_services_list - List Cloud Run services

• gcloud_run_services_describe - Get service details and URLs

Cloud Storage (2 tools)

• gcloud_storage_buckets_list - List storage buckets

• gcloud_storage_ls - List objects in buckets

IAM (2 tools)

• gcloud_iam_service_accounts_list - List service accounts

• gcloud_iam_service_accounts_keys_create - Create service account keys

Generic Wrapper (1 tool)

• gcloud_execute - Execute any gcloud command with proper ADC authentication

Prerequisites:

• gcloud CLI installed: brew install google-cloud-sdk

• Authenticated: gcloud auth application-default login

• Project set: gcloud config set project PROJECT_ID

Documentation:

• GCP_README.md - Complete overview

• GCP_QUICKSTART.md - Setup guide

• GCP_TOOLS.md - All tools reference

• GCP_EXAMPLES.md - Real-world examples

📋 Planning & Requirements (3 tools)

generate_requirements

Transform ideas into comprehensive requirements:

• Functional and non-functional requirements

• User stories with acceptance criteria

• Technical constraints and success metrics

• Monetization strategy

• Complete PRD from a prompt

create_product_roadmap

Generate development roadmap:

• Phase breakdown (MVP → Full → Enterprise)

• Milestones and timelines

• Feature prioritization

• Clear path from idea to launch

generate_user_stories

Create detailed user stories:

• Acceptance criteria in Given/When/Then format

• Priority and story point estimation

• Ready for sprint planning

🔬 Research & Analysis (3 tools)

analyze_tech_stack

Recommend optimal technologies:

• Analyze project requirements

• Recommend frameworks, databases, hosting

• Provide alternatives with reasoning

• Consider team size and expertise

• Data-driven technology decisions

research_best_practices

Industry best practices database:

• Security patterns (OWASP, authentication)

• Performance optimization techniques

• Testing strategies and patterns

• Deployment best practices

• Learn from industry leaders

competitive_analysis

Market and competitive intelligence:

• Identify opportunities and threats

• Differentiation strategies

• Market positioning recommendations

• Strategic product decisions

🏗️ Architecture & Design (3 tools)

design_system_architecture

Complete system architecture design:

• Layered architecture (Presentation, Application, Data, Infrastructure)

• Architecture patterns (microservices, event-driven, CQRS)

• Scalability and security strategies

• Data flow diagrams

• Production-ready architecture from day one

design_database_schema

Database schema design:

• Entity modeling with relationships

• SQL DDL generation

• Indexes for performance

• Migration planning

• Optimized data layer

generate_api_spec

OpenAPI/Swagger specification:

• Complete endpoint definitions

• Request/response schemas

• Authentication schemes

• Contract-first API development

🔒 Security & Compliance (3 tools)

security_audit

Comprehensive security scanning:

• Dependency vulnerability scanning (npm audit)

• Exposed secrets detection

• Code security issues (eval, SQL injection, XSS)

• Severity-based recommendations

• Find vulnerabilities before attackers do

generate_security_policy

Security policy documentation:

• Authentication/authorization guidelines

• Data protection measures

• Incident response plans

• Compliance checklists (OWASP, GDPR, SOC 2)

• Enterprise-grade security documentation

scan_for_vulnerabilities

Targeted vulnerability scanning:

• SAST (Static Application Security Testing)

• Dependency checks

• Secret scanning

• OWASP Top 10 validation

• Continuous security monitoring

🚀 Deployment & Infrastructure (3 tools)

generate_dockerfile

Optimized Docker containers:

• Multi-stage builds for minimal size

• Security best practices (non-root user)

• Language-specific optimizations

• .dockerignore generation

• Production-ready containerization

generate_cicd_pipeline

CI/CD automation:

• GitHub Actions / GitLab CI workflows

• Automated testing and building

• Security scanning in pipeline

• Deployment automation

• Zero-touch deployments

generate_kubernetes_manifests

Kubernetes deployment configs:

• Deployments with replicas

• Services and load balancers

• Ingress with TLS

• Health checks (liveness/readiness)

• Resource limits and requests

• Cloud-native deployment ready

🎓 Autonomous Development Instructions

This server includes comprehensive instructions in .github/instructions/ that guide AI assistants to:

Core Behaviors

1. Zero Human Intervention - Operate autonomously by default

2. Fix, Don't Report - Iterate until issues are resolved

3. Test Everything - No untested code ships

4. Document Thoroughly - Always up-to-date docs

5. Automate Relentlessly - Scripts for all common tasks

Quality Standards

• ✅ All tests must pass

• ✅ Zero linting errors

• ✅ Code coverage >80%

• ✅ TypeScript strict mode

• ✅ Comprehensive error handling

• ✅ Production-ready on first ship

Workflow Automation

• Auto-run tests after code changes

• Auto-fix linting issues

• Auto-update documentation

• Auto-commit with semantic messages

• Auto-generate validation scripts

See .github/instructions/autonomous-development.instructions.md for complete guidelines.

📖 Installation

git clone https://github.com/yourusername/mcp-vibe-coding-tools.git
cd mcp-vibe-coding-tools
npm install
npm run build

🔧 Configuration

For VS Code & GitHub Copilot

See VSCODE_SETUP.md for detailed setup instructions.

Quick setup: Run MCP: Open User Configuration from Command Palette and add:

{
  "servers": {
    "mcp-vibe-coding-tools": {
      "type": "stdio",
      "command": "node",
      "args": ["/path/to/mcp-vibe-coding-tools/dist/mcp-vibe-coding-tools.js"]
    },
    "mcp-gitops-tools": {
      "type": "stdio",
      "command": "node",
      "args": ["/path/to/mcp-vibe-coding-tools/dist/gitops-server.js"],
      "env": {
        "GITHUB_API_KEY": "your-github-token",
        "GITLAB_API_KEY": "your-gitlab-token",
        "GITLAB_HOST": "https://gitlab.com"
      }
    }
  }
}

Replace /path/to/mcp-vibe-coding-tools with the absolute path to where you cloned this repo. Then it works in any project you open in VS Code!

For Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "vibe-coding-tools": {
      "command": "node",
      "args": ["/path/to/mcp-vibe-coding-tools/dist/mcp-vibe-coding-tools.js"],
      "env": {
        "WORKSPACE_PATH": "/path/to/your/current/project"
      }
    },
    "gitops-tools": {
      "command": "node",
      "args": ["/path/to/mcp-vibe-coding-tools/dist/gitops-server.js"],
      "env": {
        "GITHUB_API_KEY": "your-github-token",
        "GITLAB_API_KEY": "your-gitlab-token",
        "GITLAB_HOST": "https://gitlab.com"
      }
    }
  }
}

For Claude, you may want to update WORKSPACE_PATH per project, or use a default working directory.

For Other MCP Clients

The server uses stdio transport (standard for MCP) - configure similarly in:

• Cursor

• Windsurf

• Cline

• Continue

• Any MCP-compatible client

Same pattern:

• Main server: node /path/to/server/dist/mcp-vibe-coding-tools.js with optional WORKSPACE_PATH env var

• GitOps server: node /path/to/server/dist/gitops-server.js with GITHUB_API_KEY and/or GITLAB_API_KEY env vars

🎯 Usage Examples

Autonomous Feature Development

Prompt: "Add user authentication with JWT, including tests and docs"

AI will automatically:
1. Write authentication module with error handling
2. Create comprehensive unit & integration tests
3. Run tests and fix any failures
4. Generate API documentation
5. Update README with usage examples
6. Create validation scripts
7. Set up CI/CD for auth tests
8. Commit with semantic message

You get: Production-ready, tested, documented feature ✅

Autonomous Bug Fixing

Prompt: "Fix the memory leak in data processor"

AI will automatically:
1. Analyze code to locate leak
2. Write fix with proper cleanup
3. Add regression tests
4. Run full test suite
5. Verify fix with profiling
6. Update CHANGELOG
7. Commit fix

You get: Bug fixed, tested, documented ✅

Autonomous Project Setup

Prompt: "Create a new Express API project with full automation"

AI will automatically:
1. Initialize project structure
2. Add TypeScript, testing, linting
3. Create validation scripts
4. Set up GitHub Actions
5. Add pre-commit hooks
6. Generate documentation
7. Create example endpoints with tests

You get: Production-ready project template ✅

🏗️ Architecture

Built with modern MCP SDK:

• McpServer class (not deprecated Server)

• registerTool() method (not deprecated tool())

• Stdio transport for universal compatibility

• Structured responses with proper error handling

• Type-safe with TypeScript strict mode

🔒 Security

• Path validation prevents directory traversal

• No deletion tools - files are never auto-deleted

• Sandboxed execution - workspace-scoped operations

• Input sanitization for all user data

• Environment isolation with virtual environments

🤝 Contributing

See CONTRIBUTING.md for development workflow.

This project follows autonomous development principles:

• All PRs must pass validation

• Tests required for new features

• Documentation updated automatically

• CI/CD enforces quality standards

📝 License

MIT - See LICENSE file

🙏 Acknowledgments

Built with:

• Model Context Protocol - MCP standard

• MCP TypeScript SDK - Official SDK

• Modern development best practices

🚦 Status

• ✅ 114 production tools across 2 servers

• ✅ Full autonomous workflow support

• ✅ Multi-language support (JS/TS, Python, Rust, Go)

• ✅ Comprehensive validation automation

• ✅ Self-documenting capabilities

• ✅ CI/CD integration ready

• ✅ Zero deprecated APIs

Unchain your AI development workflow. Ship production code autonomously.