Agentpack

@ihorponom

Agentpack について

Agentpack is a local-first continuity layer for AI coding agents. It stores reviewed task state inside a repo-local .agentpack/ ledger so Codex, Claude Code, Cursor, Claude Desktop, and other MCP clients can resume work without rebuilding context from chat history.

#cli-typescript-mcp-developer-tools-cursor-codex-ai-agents-agent-state-local-first-coding-agents-model-context-protocol-mcp-server-claude-code-context-engineering-task-continuity

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": {
    "agentpack": {
      "command": "npx",
      "args": [
        "-y",
        "agentpack-cli@latest",
        "mcp",
        "--root",
        "/absolute/path/to/your/repo"
      ],
      "env": {
        "AGENTPACK_ROOT": "/absolute/path/to/your/repo"
      }
    }
  }
}

ツール

ツールは検出されませんでした

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 Agentpack?

Agentpack is a local repo-scoped task ledger for MCP-connected coding agents. It lets agents record decisions, dead ends, evidence, source conclusions, task passports, checkpoints, and compact resume context so future sessions can continue without rebuilding context from chat history. It works with Codex, Claude Code, Cursor, Claude Desktop, and other MCP clients.

How to use Agentpack?

Install the CLI tool globally via npm (npm install -g agentpack-cli), then run agentpack init inside your repository. Next, run agentpack install <client> --write (e.g., agentpack install codex --write or agentpack install claude --write) to configure the integration for your preferred coding agent.

Key features of Agentpack

Records durable task decisions and failed approaches
Preserves reviewed source conclusions with file hashes
Creates checkpoints and handoffs for repo-scoped work
Supports switching between Codex, Claude Code, Cursor, and Claude Desktop
Provides compact resume context to avoid chat‑history rebuilding

Use cases of Agentpack

Continuing a coding session after chat compaction or a fresh agent session
Seamlessly switching between different MCP clients (Codex, Claude Code, Cursor, Claude Desktop)
Recording decisions and dead ends to avoid repeating work
Handing off task context between team members or agent sessions

FAQ from Agentpack

What does Agentpack record?

Agentpack records decisions, dead ends, evidence, source conclusions with file hashes, task passports, checkpoints, and compact resume context.

Which clients does Agentpack support?

Agentpack works with Codex, Claude Code, Cursor, Claude Desktop, and other MCP‑connected coding agents.

How do I install Agentpack?

Install the CLI globally with npm install -g agentpack-cli, then run agentpack init in your repository and agentpack install <client> --write for your chosen agent.

Can Agentpack help after a chat compaction?

Yes, it is explicitly designed to provide continuity after chat compaction or a fresh agent session by preserving a local task ledger.

Does Agentpack require any external services?

No, Agentpack operates as a local repo‑scoped ledger; no external service or cloud dependency is mentioned.

よくある質問

What does Agentpack record?

Agentpack records decisions, dead ends, evidence, source conclusions with file hashes, task passports, checkpoints, and compact resume context.

Which clients does Agentpack support?

Agentpack works with Codex, Claude Code, Cursor, Claude Desktop, and other MCP‑connected coding agents.

How do I install Agentpack?

Install the CLI globally with `npm install -g agentpack-cli`, then run `agentpack init` in your repository and `agentpack install <client> --write` for your chosen agent.

Can Agentpack help after a chat compaction?

Yes, it is explicitly designed to provide continuity after chat compaction or a fresh agent session by preserving a local task ledger.

Does Agentpack require any external services?

No, Agentpack operates as a local repo‑scoped ledger; no external service or cloud dependency is mentioned.

基本情報

カテゴリ

AI とエージェント

トランスポート

stdio

公開者

ihorponom

投稿者

Ihor Ponomarov

「AI とエージェント」の他のコンテンツ

Open Multi-Agent Canvas

CopilotKit

The open-source multi-agent chat interface that lets you manage multiple agents in one dynamic conversation and add MCP servers for deep research

507

MCP Client for Ollama (ollmcp)

jonigl

Harness the power of local LLMs with this TUI MCP Client for Ollama. Featuring all core MCP primitives (tools, prompts, resources), agent mode, multi-server, model switching, streaming responses, human-in-the-loop, thinking mode, model params config, system prompts, and saved pre

768