@ha-bits/bit-agent

AI Agent bit with LangChain and MCP (Model Context Protocol) integration for multi-tool AI workflows.

Features

• Multi-provider LLM support: OpenAI, Anthropic (Claude), Google (Gemini)
• MCP Integration: Connect to any MCP server for tool access
• Built-in MCP presets: Google Drive, Slack, GitHub, Notion, Figma, Everything
• Custom MCP servers: Define your own MCP server connections
• Dual transport: Support for both stdio and SSE (Server-Sent Events) transports
• Dynamic configuration: Configure MCPs dynamically in your habit YAML

Installation

npm install @ha-bits/bit-agent

Usage

Basic Example

nodes:
  - id: agent-search
    type: bit
    data:
      framework: bits
      module: '@ha-bits/bit-agent'
      operation: run_agent
      source: npm
      credentials:
        agentAuth:
          llmProvider: openai
          apiKey: '{{habits.env.OPENAI_API_KEY}}'
          mcpSecrets:
            SLACK_BOT_TOKEN: '{{habits.env.SLACK_BOT_TOKEN}}'
            GDRIVE_CREDENTIALS: '{{habits.env.GDRIVE_CREDENTIALS}}'
      params:
        prompt: 'Search for documents about Q4 planning'
        mcpServers:
          - type: google-drive
          - type: slack

With Custom MCP Server

params:
  prompt: 'Analyze the local codebase'
  mcpServers:
    - type: custom
      command: npx
      args: ['@anthropic/mcp-server-filesystem', '/path/to/project']
      transport: stdio

With SSE Transport

params:
  prompt: 'Query the remote database'
  mcpServers:
    - type: custom
      transport: sse
      url: 'http://localhost:8080/mcp'

Built-in MCP Presets

| Preset | Description | Required Secrets | |--------|-------------|------------------| | google-drive | Google Drive file access | GOOGLE_DRIVE_CREDENTIALS | | slack | Slack workspace access | SLACK_BOT_TOKEN | | github | GitHub repository access | GITHUB_TOKEN | | notion | Notion workspace access | NOTION_API_KEY | | figma | Figma design access | FIGMA_ACCESS_TOKEN | | everything | Local filesystem & testing | None |

Authentication

The credentials object must be nested under agentAuth key:

credentials:
  agentAuth:
    llmProvider: openai  # or 'anthropic', 'google'
    apiKey: '{{habits.env.OPENAI_API_KEY}}'
    mcpSecrets:
      SLACK_BOT_TOKEN: '{{habits.env.SLACK_BOT_TOKEN}}'

Action Parameters

run_agent

| Parameter | Type | Required | Default | Description | |-----------|------|----------|---------|-------------| | prompt | string | Yes | - | The task or question for the agent | | systemPrompt | string | No | - | System instructions for the agent | | mcpServers | array | Yes | - | List of MCP servers to connect | | model | string | No | Provider default | Model to use | | maxIterations | number | No | 10 | Maximum agent iterations | | temperature | number | No | 0.7 | LLM temperature |

License

MIT