SkillHub MCP is an AI resource intelligence server for discovering, searching, and getting recommendations from 20,000+ AI tools, skills, agents, rules, and MCP servers — fully offline with no runtime dependencies.
recommend: Describe a task in natural language and get ranked, relevance-scored recommendations of AI tools, agents, MCP servers, skills, and more. Supports filtering by type and ecosystem.
search: Perform targeted keyword searches across the database, filtering by resource type (skill, agent, tool, mcp-server, etc.) or ecosystem (cursor, openai, anthropic, mcp, etc.).
get_resource: Retrieve full details about a specific resource by ID or exact name.
get_setup_guide: Fetch installation and configuration instructions for any resource.
analyze_stack: Describe your technology stack and receive tailored recommendations for complementary AI tools, coding rules, agents, and MCP servers.
stats: View statistical information about the resource database.
diagnostics: Check system health, including Node.js version, database loading, and AI client configuration status.
generate_config: Produce manual configuration files for AI clients like Codex, Claude, Cursor, and Windsurf.
The server operates securely and privately — fully offline, zero runtime dependencies, no shell access, and limited filesystem access.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@SkillHub MCP recommend tools for building a RAG pipeline with LangChain"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
AI Resource Intelligence — Find the right tool for any AI task.
Search, discover, and get recommendations from 20,000+ skills, tools, agents, rules, and MCP servers — all from your terminal or AI client.
npx skillhub-mcp setupThis auto-detects your installed AI clients (Codex, Claude, Cursor, Windsurf) and configures them automatically.
Your AI assistant can now discover and recommend tools. Try asking it:
"What tools should I use to build a RAG pipeline with LangChain?"
Related MCP server: evc-spark-mcp
SkillHub also works as a standalone CLI — no MCP client required.
npx skillhub-mcp recommend "build a REST API with authentication"Output:
#1 Express.js ✓ 9.2
Fast, unopinionated web framework for Node.js
tool · cross-platform · expressjs
→ matches technologies: express, node; relevant to: api, coding
#2 Passport.js 8.8
Authentication middleware for Node.js
tool · cross-platform · jaredhanson
→ matches technologies: node; relevant to: authenticationnpx skillhub-mcp search "vector database"npx skillhub-mcp info "LangChain"npx skillhub-mcp statsAll data commands support --json for piping and scripting:
npx skillhub-mcp recommend "kubernetes" --json | jq '.results[0]'
npx skillhub-mcp stats --jsonWhen connected to an AI client, SkillHub exposes these tools:
Tool | Description |
| Analyze a task and recommend relevant resources |
| Search 20,000+ resources by keyword |
| Get full details for a specific resource |
| Get install instructions for any resource |
| Recommend tools based on your tech stack |
This package is designed to be safe and auditable.
✅ Fully offline — zero network requests, no HTTP calls, no telemetry
✅ Zero runtime dependencies — self-contained bundle, nothing from node_modules
✅ No install scripts — no postinstall, no code runs on npm install
✅ No eval — no eval(), new Function(), or dynamic code execution
✅ No shell access — no exec(), spawn(), or child_process usage
✅ Filesystem access only in setup wizard — writes MCP client configs when you explicitly run setup
See SECURITY.md for full details on runtime behavior and filesystem access.
npx skillhub-mcp doctorChecks Node.js version, npx path resolution, database loading, and MCP client configuration status.
If the setup wizard doesn't work for your client, generate the config manually:
npx skillhub-mcp print-config codex # Codex (TOML)
npx skillhub-mcp print-config claude # Claude Desktop (JSON)
npx skillhub-mcp print-config cursor # Cursor (JSON)
npx skillhub-mcp print-config windsurf # Windsurf (JSON)Important: The setup wizard uses absolute paths for
npx(e.g.,/opt/homebrew/bin/npx). GUI applications often don't inherit your shell's PATH.
Node.js 18+
npm / npx
MIT
Become a sponsor on Tribe.run.
Analyze a technology stack description and recommend complementary AI tools, skills, MCP servers, and rules that would enhance the developer's workflow. Use this when the user describes their project, tech stack, or development environment and could benefit from AI-powered tools.
| Name | Required | Description | Default |
|---|---|---|---|
| focus | No | What kind of recommendations to focus on | all |
| stack | Yes | Description of the user's technology stack (e.g., 'Next.js, TypeScript, Prisma, Tailwind CSS, deployed on Vercel') | |
| maxResults | No | Max results per category |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden. It discloses that the tool recommends specific categories (AI tools, skills, MCP servers, rules), which is adequate. However, it does not mention any behavioral traits like rate limits, authentication, or side effects, leaving gaps in transparency.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is two sentences, front-loading the action and then providing usage context. Every sentence adds value with zero fluff.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given no output schema and simple parameters, the description covers the main purpose and usage. It does not detail the return format, but for a recommendation tool, the context is reasonably complete. A slightly higher score would require output structure hints.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100%, so the baseline is 3. The description does not add significant meaning beyond the schema; it mentions the categories of recommendations which align with the 'focus' parameter enum but does not explain parameters in more detail.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool analyzes a technology stack and recommends complementary AI tools, skills, MCP servers, and rules. This distinguishes it from siblings like 'recommend' which may be broader, and 'search' which is for searching resources.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description explicitly says 'Use this when the user describes their project, tech stack, or development environment and could benefit from AI-powered tools.' This provides clear context but does not specify when not to use or mention alternatives, so it misses the full guidance for a 5.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Get full details about a specific AI resource from SkillHub by its ID or exact name. Use this when the user wants more information about a previously recommended resource.
| Name | Required | Description | Default |
|---|---|---|---|
| identifier | Yes | Resource ID or exact title to look up |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description must disclose behavioral traits. It says 'Get full details' but does not explain what 'full details' includes, whether it's read-only, error handling for invalid identifiers, or any side effects. For a simple lookup, it's adequate but not thorough.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two sentences: first states what the tool does, second states when to use it. No wasted words; front-loaded with essential information. Excellent conciseness.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's low complexity (1 parameter, no output schema, no annotations), the description is mostly complete. It covers purpose and usage context. However, it omits any detail about return format or constraints, which prevents a perfect score.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The schema covers the single parameter 'identifier' with a description ('Resource ID or exact title to look up') at 100% coverage. The description adds no new semantic value beyond restating the parameter's role, so baseline 3 applies.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
Description clearly states 'Get full details about a specific AI resource' with a specific verb and resource. It indicates lookup by ID or exact name. While it doesn't explicitly differentiate from siblings like search or recommend, the purpose is clearly about retrieving details of a known resource.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description explicitly states 'Use this when the user wants more information about a previously recommended resource,' providing clear context. It could be improved by mentioning when not to use (e.g., if the resource is already fully known or for other operations), but the guidance is direct and helpful.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Get installation and setup instructions for a specific AI resource from SkillHub. Use this after recommending a resource to help the user actually install and configure it.
| Name | Required | Description | Default |
|---|---|---|---|
| identifier | Yes | Resource ID or exact title |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden of behavioral disclosure. It describes the tool as returning instructions ('get'), implying a safe read operation, but does not explicitly state idempotency, side effects, or auth requirements. It lacks detail on whether the tool alone provides complete setup guidance.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two sentences and zero wasted words. The first sentence fronts the core purpose, and the second provides immediate usage context. Highly efficient.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a simple tool with one parameter and no output schema, the description adequately covers purpose and usage. It could optionally describe the return format, but this is not critical given the tool's simplicity and lack of output schema.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema has one parameter 'identifier' with a clear description. The tool description adds usage context but no additional semantic meaning beyond what the schema already provides. With 100% schema coverage, a score of 3 is appropriate as per guidelines.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description begins with a specific verb-resource pair: 'Get installation and setup instructions for a specific AI resource from SkillHub.' It clearly distinguishes from sibling tools like 'get_resource' (which likely returns the resource itself) and 'recommend' (for recommendations), as it provides installation context.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description explicitly states when to use this tool: 'Use this after recommending a resource to help the user actually install and configure it.' This gives clear contextual guidance, although it does not mention specific situations where the tool should not be used.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recommend AI tools, skills, MCP servers, agents, rules, and resources from the SkillHub ecosystem (20,000+ resources) based on the user's task or intent. Use this when the user could benefit from discovering relevant AI tools, needs help finding the right framework/library, or is working on a task that could be improved with specific AI resources. Returns ranked results with relevance scores and install guidance.
| Name | Required | Description | Default |
|---|---|---|---|
| task | Yes | Description of what the user is trying to do. Can be a natural language task description, a technical question, or a prompt that implies the need for AI tools. | |
| types | No | Filter by resource types. Leave empty for all types. | |
| ecosystems | No | Filter by ecosystem. Leave empty for all ecosystems. | |
| maxResults | No | Maximum number of results to return (default: 10, max: 30) |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden. It mentions that results are ranked with relevance scores and install guidance, but does not disclose any potential side effects, authentication needs, or limitations. This is adequate but not comprehensive.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is three sentences: purpose with scope, usage guidance, and output format. It is front-loaded with essential information and contains no unnecessary words.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's recommendation nature, 4 parameters, and no output schema, the description sufficiently covers what the tool does and what it returns. It could mention pagination or sorting behavior, but the current description is adequate for an AI agent.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, so the schema already explains all parameters. The description adds some nuance to the 'task' parameter (e.g., natural language, technical question, prompt), which is helpful but not essential. Overall, the description adds marginal value beyond the schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool recommends AI tools, skills, MCP servers, agents, rules, and resources from the SkillHub ecosystem based on user's task. It distinguishes itself from sibling tools like search and get_resource by focusing on discovery and ranking.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description explicitly states when to use: when the user could benefit from discovering relevant AI tools, needs help finding the right framework/library, or is working on a task improvable with AI resources. It does not explicitly state when not to use, but the context is clear enough.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Search the SkillHub database (20,000+ AI resources) by text query, type, ecosystem, or tags. Use this for targeted lookups when the user is looking for a specific tool, compares options, or wants to browse resources in a specific category.
| Name | Required | Description | Default |
|---|---|---|---|
| type | No | Filter by resource type | |
| query | Yes | Search query — tool name, technology, or keyword | |
| ecosystem | No | Filter by ecosystem | |
| maxResults | No | Max results to return (default: 15) |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description must cover behavioral traits. It does not mention read-only nature, rate limits, or behavior on empty results. The focus is on purpose, not behavior, leaving significant gaps for agent decision-making.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two sentences that are front-loaded with the action and scope. No fluff or redundancy.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
No output schema exists, yet the description does not mention return format, pagination, or error handling. It also contains an inaccuracy about tags. For a 4-parameter search tool with no annotations, this is insufficient.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100%, so baseline is 3. However, the description claims filtering by 'tags' which is not a parameter in the schema, causing potential confusion. It adds little beyond the schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the verb 'Search' and the resource 'SkillHub database (20,000+ AI resources)' with specific filtering dimensions (text query, type, ecosystem, or tags). It also explains the usage scenarios (lookups, comparisons, browsing), distinguishing it from sibling tools like get_resource or recommend.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description gives clear context for when to use the tool ('targeted lookups', 'compares options', 'browse resources in a specific category'). However, it does not explicitly mention when not to use it or provide alternatives, but the guidance is sufficient.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
v0.6.0analyze_stackget_resourceget_setup_guiderecommendsearchScored across 5 tools
Each tool has a clear, distinct purpose: analyze_stack for stack analysis, recommend for general recommendations, search for database queries, get_resource for details, and get_setup_guide for installation instructions. There is no overlap, and agents can easily distinguish them.
Most tools follow a verb_noun pattern (analyze_stack, get_resource, get_setup_guide), but recommend and search are single verbs. This is a minor inconsistency but still predictable and readable.
With 5 tools, the set is well-scoped for a resource discovery and recommendation server. Each tool serves a necessary function without redundancy or bloat.
The tools cover the core workflow: stack analysis, general recommendation, search, resource details, and setup guides. Minor gaps like user feedback or comparison features exist but do not hinder the primary purpose.
Search & install 6,500+ AI agent skills from skills-hub.ai inside any MCP tool.
A registry of 5,900+ peer-authored skills any MCP agent can search and load on demand.
AI agent registry — search, discover, register, and connect agents via MCP.
Search and install 4,000+ security-scanned MCP servers from inside any MCP-aware AI client.
