Skip to main content
Glama
Brain Bridge
by 0xlicy

✨ Features

šŸ”Œ MCP Server

  • Cross-conversation artifact access — Read, list, and search artifacts across all Antigravity conversations

  • Shared notes — Write notes accessible by any conversation (_shared/ directory)

  • Task delegation — Spawn headless child conversations to parallelize complex work

  • Task monitoring — Check status, results, and manage delegated sub-tasks

šŸ–„ļø VS Code Extension

  • Sidebar UI — Dedicated activity bar panel with three views:

    • šŸ“‹ Delegated Tasks — Real-time task status with inline actions (check result, focus, cancel)

    • šŸ’¬ Conversations — Browse all Antigravity conversations and their artifacts

    • šŸ“ Shared Notes — View notes written by any conversation

  • One-click delegation — Delegate tasks from the command palette or status bar

  • Auto-refresh — Tasks update automatically every 5 seconds

  • File watcher — Instant pickup of MCP-delegated tasks

Related MCP server: kitty-hive

šŸ—ļø Architecture

brain-bridge/
ā”œā”€ā”€ src/                    # MCP Server (TypeScript)
│   └── index.ts            # Server with 8 MCP tools
ā”œā”€ā”€ ext/                    # VS Code Extension
│   └── src/
│       ā”œā”€ā”€ extension.ts        # Extension entry point
│       ā”œā”€ā”€ task-delegation.ts  # Task delegation manager
│       └── task-tree-provider.ts # TreeView data providers
ā”œā”€ā”€ .agents/                # Agent skills
│   └── skills/parallel-delegation/
ā”œā”€ā”€ test/                   # Test suite
└── package.json

Data directory: ~/.gemini/antigravity/brain/

  • <uuid>/ — Conversation artifacts (task.md, implementation_plan.md, etc.)

  • _shared/ — Cross-conversation shared notes

  • _tasks/ — Delegated task state files

šŸš€ Quick Start

Prerequisites

  • Node.js ≄ 18

  • Antigravity (or any MCP-compatible IDE)

1. Install the MCP Server

# Clone the repository
git clone https://github.com/0xlicy/antigravity-brain-bridge.git
cd antigravity-brain-bridge

# Install dependencies & build
npm install
npm run build

2. Configure MCP

Add the following to your MCP client configuration (e.g., Antigravity settings):

{
  "mcpServers": {
    "brain-bridge": {
      "command": "node",
      "args": ["/absolute/path/to/antigravity-brain-bridge/dist/index.js"]
    }
  }
}

Tip: Replace /absolute/path/to/ with the actual path where you cloned the repo.

3. Install the VS Code Extension (Optional)

cd ext

# Install dependencies & build
npm install
npm run build

# Package and install
npm run package
# Then install the generated .vsix file in Antigravity / VS Code

4. Verify

Once configured, you can use the MCP tools in any conversation:

# List all conversations
→ list_conversations

# Search across all artifacts
→ search_artifacts(query: "authentication")

# Delegate a task to a child conversation
→ delegate_task(task: "Write unit tests for the auth module")

šŸ“– MCP Tools Reference

Tool

Description

list_conversations

List all Antigravity conversations with summaries

list_artifacts

List artifact files in a specific conversation

read_artifact

Read the full content of a conversation artifact

search_artifacts

Full-text search across all artifacts

write_shared_note

Write a note to the shared _shared/ directory

delegate_task

Delegate a task to a new child conversation

check_task_result

Check the status and result of a delegated task

list_delegated_tasks

List all delegated tasks with optional status filter

šŸ› ļø Development

# Build the MCP server
npm run build

# Run the MCP server
npm start

# Run tests
npm test

Extension development:

cd ext

# Watch mode (rebuild on changes)
npm run dev

# Package into .vsix
npm run package

šŸ¤ Contributing

  1. Fork the repository

  2. Create your feature branch (git checkout -b feature/amazing-feature)

  3. Commit your changes (git commit -m 'Add amazing feature')

  4. Push to the branch (git push origin feature/amazing-feature)

  5. Open a Pull Request

šŸ“„ License

This project is licensed under the MIT License — see the LICENSE file for details.

Available Tools

8 tools
check_task_resultA

Check the status and result of a delegated task. Returns the task description, current status, and result if completed.

Inputs: task_id (required).

delegate_taskA

Delegate a task to a new child conversation. Creates a pending task that the Brain Bridge Extension will automatically pick up and execute via a headless cascade. Use check_task_result to monitor progress.

Inputs: task (required), cwd.

list_artifactsA

List all artifact files in a specific conversation. Returns the markdown files with their sizes and metadata summaries.

Inputs: conversation_id (required).

list_conversationsA

List all Antigravity conversations with their summaries. Scans the brain directory for conversation folders and reads metadata to provide summaries.

Inputs: limit.

list_delegated_tasksB

List all delegated tasks and their current status. Optionally filter by status.

Inputs: status.

read_artifactA

Read the full content of a specific artifact file from a conversation.

Inputs: filename (required), conversation_id (required).

search_artifactsA

Full-text search across artifact files. Searches all conversations or a specific one. Returns matching lines with context.

Inputs: query (required), conversation_id.

write_shared_noteA

Write a shared note to the brain/_shared/ directory, accessible across all conversations.

Inputs: title (required), content (required).

Tool Schema Changelog

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

Loading tool schema changelog…

    TDQS

    A3.9/5.0

    Scored across 8 tools

    Disambiguation5/5

    Each tool targets a distinct resource and action: conversations, artifacts, shared notes, and delegated tasks. No two tools overlap in purpose, making selection unambiguous.

    Naming Consistency5/5

    All tools follow a consistent verb_noun pattern (e.g., list_conversations, read_artifact, delegate_task). This predictability aids agent comprehension and selection.

    Tool Count5/5

    With 8 tools, the server is well-scoped. It covers the core domains (conversations, artifacts, notes, tasks) without unnecessary bloat or sparse coverage.

    Completeness3/5

    The tool set covers list/read/search for artifacts and task delegation, but lacks conversation detail reading, artifact writing, shared note reading, and task cancellation. These gaps create potential dead ends (e.g., writing a note with no way to read it).

    Related MCP Connectors

    Related MCP Servers