# docker-agent を Homebrew, winget でインストール
docker-agent のインストール経路、実行ファイル、メタデータ、AI エージェント向けセキュリティノートを確認します。
## インストール
```sh
sudo av install brew:docker-agent
```
追加のインストールコマンド:
### macOS
- Homebrew (100%):
```sh
brew install docker-agent
```
証拠: local Homebrew formula metadata
### Windows
- winget (92%):
```sh
winget install --id Docker.Agent -e
```
証拠: Windows Package Manager source index: Docker.Agent from https://cdn.winget.microsoft.com/cache/source.msix
## パッケージ情報
- **パッケージキー:** brew:docker-agent
- **パッケージマネージャ:** Homebrew
- **バージョン:** 1.120.0
- **ソース概要:** Agent Builder and Runtime by Docker Engineering
- **ホームページ:**
- **リポジトリ:**
- **最終更新:** 2026-08-03T15:08:48Z
- **生成日時:** 2026-08-03T19:37:03+00:00
## 実行可能ファイル
- docker-agent (エイリアス)
## インストール挙動
- Bottle: 利用不可
## バージョンと鮮度
- ページ生成日: 2026-08-03
- マネージャ版: 1.120.0
## プロジェクトの歴史と使われ方
Docker Agent is Docker's open source CLI for defining, running, and sharing AI agents from YAML or HCL configuration. The upstream README describes it as a docker CLI plugin that can run as docker agent, with multi-agent orchestration, MCP toolsets, provider-agnostic model support, RAG, and OCI registry distribution.
### プロジェクトの歴史
Docker Agent is a recent Docker Engineering project that applies Docker packaging habits to AI-agent workflows: declarative configuration, a CLI entry point, reusable toolsets, and distribution through registries. Its docs position it explicitly as an agent runner rather than a hosted SaaS or code framework.
### 採用の歴史
The README documents three adoption channels: preinstallation in Docker Desktop 4.63 and later, Homebrew installation, and binary releases from GitHub. That makes it unusual among early AI-agent CLIs because Docker is trying to make the agent runtime feel like another Docker workflow rather than a separate Python or Node stack.
### 使われ方
Users write an agent.yaml file with model, instructions, and toolsets, then run it with docker agent run or docker-agent directly. The docs and README highlight coding agents, operations and SRE tasks, multi-agent teams, filesystem and shell tools, MCP integrations, local models through Docker Model Runner, and pushing or pulling agents through OCI registries.
### パッケージ好きにとっての重要性
For package-manager users, Docker Agent is notable because it packages a fast-moving AI-agent runtime as a CLI plugin in the Docker ecosystem. Its significance is still emerging, but it is a clear example of AI tooling being distributed through the same channels as container developer tools.
### タイムライン
- 2026: Upstream docs describe Docker Agent as a Docker tool for running AI agents from configuration.
- 2026: The README documents Docker Desktop 4.63+ bundling, Homebrew installation, and GitHub binary releases.
- 2026: The batch input lists Homebrew and winget package-manager entries.
### Related projects
- Docker Agent overlaps with Docker Model Runner, Docker's MCP catalog, general MCP servers, local-model runtimes, and AI coding-agent CLIs, but its official framing is Docker-native agent configuration and distribution.
### ソース
-
-
-
## セキュリティノート
generalized runtime or code generation signal.
- **Geiger リスク:** yellow / 中
- generalized runtime or code generation signal
## Configuration and credential file locations
These source-backed paths show where this package keeps local settings or durable credentials. Automic Vault can use them as review targets for secret scanning, migration, and command approval.
## Configuration files
- Unix: ./agent.yaml
## 他のパッケージマネージャ記録
- winget - Docker.Agent: normalized package name match | Windows Package Manager source index: Docker.Agent from https://cdn.winget.microsoft.com/cache/source.msix
## Combined YAML source
View the package source record on GitHub. [combined/docker-agent.yml](https://github.com/mxcl/pkgdb/blob/main/combined/docker-agent.yml)
## ソース
- pkg.so package database
- Geiger risk classifier
- curated configuration and credential file locations
- curated package history
- pkgdb category and tag curation
- external package-manager database matches
- cross-ecosystem install command graph