AgentHubAgentHub

Use cases/AI knowledge base

AI knowledge & docs MCP

Ground AI with up-to-date library docs and wikis—fewer hallucinations when coding.

1,898 resources matched · Page 1 of 40

What is AI knowledge base?

Knowledge-base MCPs fix the model's biggest weaknesses—stale knowledge and hallucinations. Context7 fetches versioned official docs and code examples on demand; DeepWiki compiles any GitHub repo into a queryable wiki; vector-store and knowledge-graph servers ground agents in your own content.

Typical coding loop: ask 'What is fetch caching in Next.js 15?' inside Cursor → Context7 returns docs for that exact version → generated code stops using outdated APIs. Especially valuable during major upgrades and framework evaluation.

Before wiring in an internal wiki, scope access: give the agent a read-only, space-permissioned retrieval path instead of indexing every confidential doc, and mind licensing when exposing external content.

Good for

  • Latest API docs
  • Framework upgrade Q&A
  • Team wiki Q&A
  • Tech research

Not ideal for

  • Replacing careful reading of official docs
  • Secrets without access control
Common stack: Context7 + GitHub MCP → docs plus repo cross-check

Editor's picks

Related resources

Browse all →

Notion

v1.2.4

SkillClawHub11.5k

Notion API integration with managed OAuth. Query databases, search pages, and read workspace content. Write operations require explicit user confirmation of the target resource and connection. Use this skill when users want to interact with Notion workspaces, databases, or pages. For other third party apps, use the api-gateway skill (https://clawhub.ai/byungkyu/api-gateway). Calls run through the `maton` CLI with OAuth login, or over raw HTTP with a Maton API key where the CLI cannot be installed. Every call is authenticated as the user's connection and reaches only what that connection's authorization allows, which the provider enforces on every request; the endpoints documented here are the ones this skill uses, and any other endpoint of this app needs the user to ask for it by name. Default to read and list calls, and confirm every write or new connection with the user. This file also documents the three constructs that turn a Notion connection into automation, in the order they are used: the connection (t

source

Performs web searches using DuckDuckGo to retrieve real-time information from the internet. Use when the user needs to search for current events, documentation, tutorials, or any information that requires web search capabilities.

v1.0.0

SkillClawHub31.6k

Performs web searches using DuckDuckGo to retrieve real-time information from the internet. Use when the user needs to search for current events, documentation, tutorials, or any information that requires web search capabilities.

source

google-docs

v1.2.4

SkillClawHub9.5k

Google Docs API integration with managed OAuth. Create documents, insert text, apply formatting, and manage content. Use this skill when users want to interact with Google Docs. For other third party apps, use the api-gateway skill (https://clawhub.ai/byungkyu/api-gateway). Calls run through the `maton` CLI with OAuth login, or over raw HTTP with a Maton API key where the CLI cannot be installed. Every call is authenticated as the user's connection and reaches only what that connection's authorization allows, which the provider enforces on every request; the endpoints documented here are the ones this skill uses, and any other endpoint of this app needs the user to ask for it by name. Default to read and list calls, and confirm every write or new connection with the user. This file also documents the three constructs that turn a Google Docs connection into automation, in the order they are used: the connection (the first step), a hosted function that runs a Google Docs action through the Maton SDK, and a trig

source

FAQ

How is Context7 different from searching the docs site?

Context7 resolves library + version to precise doc chunks structured for the model—no scraping, tighter context. Plain search still suits open-ended research.

How do I connect internal documentation?

Three common paths: a vector-store MCP (Qdrant / Chroma / pgvector) for semantic retrieval, official Notion / Confluence MCPs, or a self-hosted RAG service wrapped as MCP—pick by sensitivity.

Do knowledge MCPs slow responses?

One extra tool round-trip (typically a few hundred ms) in exchange for far fewer wrong-code retries. Keep snippets small and mount only needed corpora.

Setup guides for this scenario

Related scenarios