Skip to main content
Glama

hivemind-mcp

MCP server for collective debugging knowledge + project-specific knowledge bases.

What is Hivemind?

Hivemind provides two knowledge layers:

1. Public Knowledge Base (16k+ solutions)

  • Error fixes and troubleshooting from the community

  • 223+ reusable skills and workflows

  • Success-ranked solutions that improve over time

  • Think Stack Overflow for AI agents

2. Project Knowledge Bases (Your Private Hive)

  • Auto-scans your project on setup

  • Builds foundational knowledge (tech stack, architecture, database, build system)

  • Stores project-specific solutions as you work

  • Cloud storage: syncs everywhere + 10x rate limits (1000/hour)

  • Local storage: stays private on your machine (100/hour)

Related MCP server: MCP AI POC

How It Works

Public KB:

AI hits error → Search hivemind → Get ranked solutions → Report outcome

Project KB (Hive):

"create a new hive" → Auto-scan project → Store solutions as you work → Search your private knowledge

When you solve a problem, it's automatically added to your project's hive. Next session, Claude already knows how your project works.

Installation

npm install hivemind-mcp

Setup

Claude Code

claude mcp add hivemind -- npx hivemind-mcp@latest

Restart Claude Code to load the tools.

Cursor / Windsurf / Other MCP Clients

Add to your MCP config:

{
  "mcpServers": {
    "hivemind": {
      "command": "npx",
      "args": ["hivemind-mcp@latest"]
    }
  }
}

Quick Start

Tell Claude:

"create a new hive"

Claude will:

  1. Ask if you want cloud or local storage

  2. Auto-scan your project (tech stack, architecture, database)

  3. Create 5 foundational knowledge entries

  4. Give you a user_id (save this!)

That's it. Now as you work, solutions get stored in your project's hive automatically.

Using Public Knowledge

No setup needed. Just use:

  • search_kb("your error message") - Search 16k+ solutions

  • search_skills("topic") - Find reusable workflows

  • contribute_solution(...) - Share what you learned

Tools

Public Knowledge Base

search_kb(query) Search 16k+ error solutions and fixes.

search_kb("Cannot find module 'express'")
// Returns: npm install express (92% success rate)

search_skills(query, max_results?) Search 223+ reusable skills and workflows.

search_skills("deployment")
// Returns: Top 20 deployment-related skills

get_skill(skill_id) Load full details of a specific skill.

get_skill(19417)
// Returns: Complete skill instructions

count_skills() Get total number of skills in database.

count_skills()
// Returns: { total: 223 }

contribute_solution(query, solution, category?) Share a fix you discovered with the community.

contribute_solution(
  "ECONNREFUSED 127.0.0.1:5432",
  "Start PostgreSQL: brew services start postgresql",
  "database"
)

report_outcome(solution_id, outcome) Report if a solution worked. Improves rankings.

report_outcome(123, "success")  // or "failure"

Project Knowledge Base (Hive)

init_hive(project_id, project_name, storage_choice?, project_path?) Initialize your project's knowledge base with auto-scanning.

// Step 1: Get options
init_hive("my-app", "My App")
// Returns: storage options (cloud vs local)

// Step 2: Initialize with choice
init_hive("my-app", "My App", "cloud", "/path/to/project")
// Returns: user_id + confirmation (scans project automatically)

contribute_project(user_id, project_id, query, solution, category?, is_public?) Add knowledge to your project hive.

contribute_project(
  "your-user-id",
  "my-app",
  "How to deploy this project?",
  "Run: npm run build && npm run deploy",
  "deployment",
  false  // private
)

search_project(user_id, query, project_id?, include_public?) Search your project's knowledge base.

search_project(
  "your-user-id",
  "database schema",
  "my-app"
)
// Returns: Your project-specific knowledge

Features

✅ 16k+ community solutions - Ranked by success rate ✅ 223+ reusable skills - Workflows and procedures ✅ Auto-scanning - Detects tech stack, architecture, database on setup ✅ Cloud sync - 10x rate limits (1000/hour) + access everywhere ✅ Private by default - Your project knowledge stays yours ✅ FTS search - Fast full-text search across solutions ✅ Success tracking - Solutions improve based on feedback

License

MIT

Available Tools

14 tools
contribute_projectA

Add knowledge to project hive. TRIGGERS: 'add to hive', 'update hive', 'contribute to hive', 'store in hive'. When user says 'update hive', analyze recent work and contribute automatically. When user says 'add to hive', ask what they want to store. Stores solutions, patterns, pitfalls, architecture decisions, etc. Private by default, optionally public. Categories are dynamic - user can create any category name.

Inputs: project_id (required), query (required), solution (required), user_id, +3 more.

contribute_solutionB

Submit a new solution to the hivemind knowledge base.

Inputs: query (required), solution (required), category.

count_skillsA

Get total count of skills in the database.

Inputs: no named parameters.

delete_hiveA

Delete project hive and all associated knowledge entries. Use this to start fresh or remove a project's knowledge base completely.

Inputs: project_id (required), user_id, project_path.

get_hive_overviewA

Get overview of project hive including total entries, category breakdown, and recent additions. Use when user says 'show me my hive' or 'hive overview'.

Inputs: project_id (required), user_id, project_path.

get_skillA

Get detailed information about a specific skill including full instructions and executable steps.

Inputs: skill_id (required).

init_hiveA

Create a new project hive (knowledge base). TRIGGERS: 'create a new hive', 'start a hive', 'initialize hive'. Onboarding flow: First call checks if user ever used Hivemind before. If first time, asks 'Is this your first time using Claude Code?' If yes, creates CLAUDE.md with starter config. Then guides through storage choice (cloud/local). If no project_path provided, creates empty hive with starter categories. IMPORTANT: Display the 'message' field to the user EXACTLY as returned - do not condense, reformat, or summarize it.

Inputs: project_id (required), project_name (required), is_first_time_user, storage_choice, +1 more.

init_project_kbA

Initialize a project-specific knowledge base with cloud storage. Returns user_id to store for future contributions. Cloud storage users get 10x rate limits (1000/hour vs 100/hour).

Inputs: project_id (required), project_name (required), storage_type.

list_my_hivesA

List all project hives for a user. TRIGGERS: 'show me my hives', 'list my hives', 'what hives do I have', 'all my hives'. Returns project_id, project_name, and entry count for each hive. For local storage, searches current directory for .user_id files.

Inputs: user_id (required), project_path.

report_outcomeA

Report whether a solution worked or not. Helps improve solution rankings.

Inputs: outcome (required), solution_id.

search_kbA

Search the hivemind knowledge base for troubleshooting solutions, error fixes, and best practices. Returns ranked solutions with success rates.

Inputs: query (required).

search_projectB

Search project hive for knowledge. TRIGGERS: 'search my hive for [topic]', 'search hive [query]', 'find in hive [topic]', 'what does my hive know about [topic]'. Searches your private entries + optionally public entries. Returns relevant solutions, patterns, architecture decisions, etc.

Inputs: query (required), user_id, project_id, include_public, +1 more.

search_skillsA

Search for skills by topic/keyword. Returns lightweight summaries - use get_skill() for full details.

Inputs: query (required).

update_project_entryB

Update an existing project hive entry. Can edit query, solution, or category. Only works for project entries (not global hivemind KB).

Inputs: entry_id (required), user_id, query, solution, +2 more.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.5/5.0

    Scored across 14 tools

    Disambiguation3/5

    There is significant overlap between tools, particularly contribute_project and contribute_solution both adding knowledge, and search_kb, search_project, and search_skills all performing searches with unclear boundaries. However, descriptions provide some differentiation, such as project-specific vs. global scope and entry types.

    Naming Consistency4/5

    Most tools follow a consistent verb_noun pattern (e.g., contribute_project, get_hive_overview, search_kb), with minor deviations like init_hive vs. init_project_kb and list_my_hives using 'my' inconsistently. Overall, naming is readable and predictable.

    Tool Count5/5

    With 14 tools, the count is well-scoped for managing project hives and a global knowledge base, covering creation, contribution, search, and deletion. Each tool appears to serve a distinct operational need without being excessive.

    Completeness4/5

    The toolset provides good coverage for hive lifecycle (init, contribute, search, delete) and knowledge management, with minor gaps such as no explicit tool for listing all entries in a hive or managing user settings. Core workflows are supported, allowing agents to perform most expected tasks.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • F
      license
      Not graded
      quality
      D
      maintenance
      Provides instant access to a searchable knowledge base of 16,000+ community-driven troubleshooting solutions for common coding problems. Includes community feedback and smart ranking to help AI assistants find the most effective solutions.
      -
    • A
      license
      Not graded
      quality
      D
      maintenance
      Provides AI-powered development tools including code generation, refactoring, debugging, performance optimization, and test generation, along with smart prompts for code analysis and documentation, and a built-in knowledge base of coding best practices.
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      Provides AI coding tools with a shared knowledge base of real-world fixes and gotchas, verified by usage, to stop debugging errors others already solved.
      1
      MIT