Neura Market
mybolide/mcp-probe-kit logo

mybolide/mcp-probe-kit

Free

规格驱动的通用开发工作流 MCP（30 工具）：`start_feature` / `start_bugfix` / `start_ui` 编排全流程，自动 Skill + AGENTS.md bootstrap，跨项目向量记忆（Qdrant），GitNexus 代码图谱，规格校验与质量闸门。任意 MCP 客户端可用。

FreeFree tier
Type
Open Source
Visit Website GitHub (40 stars)

About mybolide/mcp-probe-kit

mcp-probe-kit is a protocol-level MCP toolkit that provides 30 tools covering the entire development lifecycle from requirements analysis to final release. It offers workflow orchestration (start_feature, start_bugfix, start_onboard, etc.), code analysis, Git tools, code generation, project management, UI/UX utilities, cross-project vector memory (Qdrant), and quality constraints that enforce code standards. Designed for developers using AI agents such as Cursor, Claude Desktop, Cline, and Continue, it helps AI understand project context through introspection, context hydration, and delegated orchestration. The toolkit is open-source, free, and includes multilingual documentation.

Key Features

30 tools covering requirements analysis to final release
Workflow orchestration (start_feature, start_bugfix, start_onboard, start_ui, start_product, start_ralph)
Code analysis and review (code_review, code_insight, fix_bug, refactor)
Git commit generation and work reports (gencommit, git_work_report)
Project initialization and specification validation (init_project, add_feature, check_spec, estimate, etc.)
UI/UX design system synchronization (ui_design_system, ui_search, sync_ui_data)
Cross-project vector memory using Qdrant and Nomic Embed (6 memory tools)
Quality constraints with single source of truth (code limits: file ≤500 lines, function ≤50 lines, nesting ≤4, parameters ≤3)

Pros & Cons

Pros
  • Context-aware: helps AI understand project intent and history
  • Comprehensive toolset covering entire development lifecycle
  • Supports all major MCP clients (Cursor, Claude Desktop, Cline, Continue)
  • Quality constraints enforce consistent code standards across all tools
  • Free and open-source with no usage limits
  • Multilingual documentation available
Cons
  • Requires Docker for local memory stack (Qdrant + Nomic Embed) if using memory features
  • Only works with MCP-compatible AI clients
  • May have a learning curve for customizing workflows and integrating with existing projects
  • Some features require specific environment setup (e.g., Docker, Git)

Best For

AI-assisted feature development and bug fixingCode review and refactoring workflowsProject onboarding for new developersUI design system management and synchronizationCross-project knowledge reuse via memory systemQuality enforcement in code generation and reviewAutomated Git commit message generation

FAQ

How do I get started with mcp-probe-kit?
Run `npx -y mcp-probe-kit@latest` and follow the quick start guide at https://mcp-probe-kit.bytezonex.com. The setup typically takes about 5 minutes.
Is mcp-probe-kit free to use?
Yes, mcp-probe-kit is completely free and open-source. There is no paid tier or usage limit.