MCP.so
ログイン
V

Vigil — Cognitive Infrastructure for AI Agents MCP Server

@AlexlaGuardia

基本情報

カテゴリ

クラウドとインフラ

トランスポート

stdio

作者

AlexlaGuardia

投稿者

Alex la Guardia

Vigil — Cognitive Infrastructure for AI Agents MCP Server について

A nervous system for AI agents — awareness daemon, frame-based tool filtering, signal protocol, session handoff, and event triggers. 14 modules, 268 tests, zero dependencies. MCP server with stdio and SSE transports.

#ai-agents#awareness#cognitive#signals#handoff#frames#daemon#multi-agent#mcp#python

Vigil — Cognitive Infrastructure for AI Agents MCP Server の設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

{
  "mcpServers": {
    "vigil": {
      "command": "vigil",
      "args": [
        "serve",
        "--transport",
        "stdio"
      ]
    }
  }
}

Vigil — Cognitive Infrastructure for AI Agents MCP Server のツール

ツール一覧はまだありません

この掲載には表示できるツール一覧がまだありません。サーバーにツールがないという意味ではありません。対応機能はプロジェクトのドキュメントをご確認ください。

ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。

概要

What is Vigil?

Vigil is a cognitive infrastructure for AI agents that coordinates awareness across agents and sessions. It is not a memory store but a nervous system that helps agents boot with full context, filter tools to save tokens, and communicate via signals. It is built for developers using agent frameworks like Claude Code, Cursor, and Claude Desktop.

How to use Vigil?

Install via pip install vigil-agent, then run vigil quickstart to set up configuration and vigil serve --transport stdio to start the MCP server. Vigil exposes 12 MCP tools over stdio or SSE, and integrates with clients such as Claude Code, Cursor, and Claude Desktop.

Key features of Vigil

  • Awareness Daemon compiles system state every 90 seconds into hot context
  • Frame-based tool filtering saves 75–85% of tokens
  • Signal Protocol enables agent coordination via lightweight event bus
  • Session Handoff preserves context with structured summaries
  • Knowledge Auto-Extract identifies recurring patterns from signals
  • Event Triggers fire webhooks, Slack alerts, or focus items

Use cases of Vigil

  • Prevent agents from starting cold by providing full awareness in <1 second
  • Reduce token waste by loading only relevant tools per context mode
  • Coordinate multiple agents without requiring them to be in the same process
  • Automatically extract and persist recurring knowledge from agent signals
  • Trigger external actions (Slack, webhooks) based on signal patterns

よくある質問

1

What runtime dependencies does Vigil require?

Vigil has zero dependencies beyond the Python standard library; MCP is optional and only needed for MCP transport.

2

Which transports and auth does Vigil support?

Vigil supports stdio and SSE transports. The README does not mention authentication; it is not described.

3

Where does Vigil store data?

Vigil uses SQLite for storage and requires zero external infrastructure.

4

How is Vigil different from memory stores?

Vigil is not a memory store; it coordinates awareness across agents and sessions, unlike tools that simply store and retrieve data.

5

What integrations are available?

Vigil integrates with Claude Code, Cursor, Claude Desktop, GitHub Actions, Slack, and Discord.

コメント

「クラウドとインフラ」の他のコンテンツ