Docker Agent
Build, run, and share AI agents with declarative YAML, a rich tool ecosystem, and multi-agent orchestration.
- Source repo
- docker/docker-agent
- Stars
- ★ 4.6k
- Last updated
- 1d ago
- License
- Apache-2.0
- Primary language
- Go
- FA score
- 46/100 · Major gaps
At a glance
- How it runs
- Works with
- Universal · cross-platformOpenAI API · Claude API
- Cost
- Free software; you pay for model usage
- Setup effort
- Low · running in minutes
- You'll need
- Typical use
- A developer who wants to quickly create a no-code AI assistant locally by writing a YAML file and running
docker agent run. - Not a fit if
- Users who don't want to install Docker or Docker Desktop
- People who prefer a GUI over the command line
- Source review
- 46/100 · Major gaps 2 safety controls not found
What does this agent do, and when should you use it?
Docker Agent is a Docker CLI plugin that lets developers define and run AI agents using simple YAML configuration, eliminating the need for coding. It supports a multi-agent architecture where specialized agents can automatically delegate tasks, and includes built-in reasoning tools like think, todo, and memory. Agents can leverage built-in tools and any MCP server (local, remote, or Docker-based), and are compatible with multiple model providers including OpenAI, Anthropic, Gemini, AWS Bedrock, Mistral, xAI, and Docker Model Runner. Agents can be packaged and pushed to any OCI registry, enabling sharing and deployment across environments. The project offers comprehensive documentation, examples, and an interactive generation tool, along with RAG capabilities (BM25, embeddings, hybrid search, reranking).
Docker Agent runs as a Docker CLI plugin, invoked via docker agent. It reads YAML configuration files (e.g., agent.yaml) that define an agent's model, instructions, and toolset, then executes the agent to perform tasks. Specific operations include: pulling and running agents from OCI registries; using docker agent new to interactively generate new agent configs; calling MCP tools (like DuckDuckGo) for web searches; leveraging built-in think, todo, and memory tools for enhanced reasoning; and supporting RAG retrieval (BM25, embeddings, etc.). Agents output results to the terminal and can be interacted with via TUI or CLI.
- A developer who wants to quickly create a no-code AI assistant locally by writing a YAML file and running
docker agent run. - A team that uses the multi-agent architecture to break a complex task into specialized agents (e.g., code review, documentation) that collaborate automatically.
- An ops engineer who runs open-source models locally via Docker Model Runner to avoid cloud API dependencies and maintain data privacy.
- A developer who packages an agent as an OCI image, pushes it to a private registry, and reuses it in CI/CD pipelines.
- A user who extends an agent's capabilities by integrating external services like search engines or databases through MCP tools.
- A knowledge worker who leverages RAG to give the agent access to an enterprise knowledge base for more accurate answers.
How do you install or deploy this agent?
Install Docker Agent via one of three methods:
- Docker Desktop 4.63+: pre-installed, just run
docker agent. - Homebrew:
brew install docker-agent, then either rundocker-agentdirectly or symlink it to~/.docker/cli-plugins/docker-agentto usedocker agent. - Download a binary from GitHub Releases and symlink it to
~/.docker/cli-plugins/docker-agent.
You must also set at least one API key, e.g., export OPENAI_API_KEY=sk-..., or use Docker Model Runner for local models.
How do you use this agent?
After installation and API key configuration, run commands like:
- docker agent run to run the default agent.
- docker agent run myorg/agent:tag to run an agent from an OCI registry.
- docker agent new to interactively generate a new agent.
- docker agent run agent.yaml to run your own YAML config.
Refer to the official quickstart guide for a full walkthrough.
What are this agent's strengths and limitations?
- Declarative YAML configuration enables defining agents without coding, making them easy to version and share.
- Multi-agent architecture with automatic delegation supports complex collaborative workflows.
- Model-agnostic: works with OpenAI, Anthropic, Gemini, Bedrock, and more, and integrates with Docker Model Runner for local models.
- MCP ecosystem support allows connecting to any MCP server, greatly extending capabilities.
- OCI registry packaging enables easy distribution and reuse across teams and environments.
- Requires Docker or Docker Desktop, which may be a barrier for non-Docker users.
- Depends on external API keys (e.g., OpenAI) unless using local models, incurring API costs.
- As a young project, community support and ecosystem maturity may be lower compared to established frameworks.
- Documentation and examples focus on CLI usage, with limited integration with IDEs or other platforms.
How does this agent compare with similar options?
Key facts side by side with the most closely related agents.
| Agent | Source review | Form / cost | Stars | Updated | Language | Full support on |
|---|---|---|---|---|---|---|
| Docker Agent This agent | 46 · Major gaps | CLIFree + model costs | ★ 4.6k | 1d ago | Go | OpenAI API · Claude API |
| RepoBrain – Repo Intelligence Engine | 49 · Major gaps | Agent plugin / skillFree + model costs | ★ 1.4k | 3d ago | Python | Codex · Claude Code |
| Agentica | 63 · Some gaps | CLIFree + model costs | ★ 356 | today | Python | OpenAI API · Claude API |
| ZhikunCode | 56 · Major gaps | Self-hosted serviceFree + model costs | ★ 525 | 1d ago | Java | OpenAI API · Claude API |
How does FollowAgents rate this agent?
Why each dimension lost points
Evidence: README mentions telemetry collecting anonymous usage data but lacks specifics on data flow and user consent; dependencies include keyring and AWS SDK suggesting sensitive credential handling, but no detailed security handling; CI workflow uses SHA-pinned actions showing dependency security practices; SECURITY.md provides vulnerability reporting process. Deductions: no user confirmation mechanism, insufficient data flow transparency, missing sensitive data handling details, no rollback mechanism mentioned, publisher identity unverified.
Evidence: README and CI workflow are consistent, build and test processes clear; dependency list complete but no availability guarantees; failure messages: CI has failure log steps, but product error handling not detailed. Deductions: insufficient evidence for failure messages.
Evidence: README describes multiple usage scenarios (local, OCI registry, interactive generation), supports multiple model providers, good environment fit; capability boundaries documented, but trigger precision (e.g., command arguments) not detailed. Deductions: insufficient evidence for trigger precision.
Evidence: README structure clear, installation instructions detailed, examples and documentation links provided; Apache-2.0 license explicit; maintenance responsibility by Docker, but no changelog. Deductions: missing versioning/changelog, known limitations not explicit.
Evidence: High output usability with CLI and TUI, rich examples; high marginal value with multi-agent and RAG features; cost-benefit: requires API keys but no cost estimates. Deductions: insufficient evidence for cost-benefit.
Evidence: Feature claims in README supported by documentation links, but no independent verification; cross-source corroboration limited, mainly single repository; facts and inferences not clearly separated. Deductions: insufficient cross-source corroboration, unclear fact-inference separation.
- Not found in source: confirmation before actingTurn on (or add) a confirmation step before it acts, and try it in a sandbox or test environment before real data.
- Not found in source: rollback or recovery pathBack up first, or work on a git branch or snapshot, so its changes can be undone.
- Publisher identity unverified; be cautious about supply chain risks.
- Telemetry data collection lacks transparency; users should understand data flow.
- Sensitive data handling details insufficient; review key storage and transmission.
- No rollback mechanism; back up configurations before upgrading.
Recent releases
Latest entries from the project's GitHub Releases. Notes are quoted from upstream as written, not rewritten or tested by FollowAgents.
-
Release notes excerpt
This release adds new built-in toolsets, provider support, and agent execution capability reporting, along with several bug fixes for runtime configuration propagation and TUI startup behavior. What's New - Adds DaoXE as a built-in OpenAI-compatible provider alias, authenticated via DAOXE_API_KEY with the api.daoxe.com/v1 base URL - Reports program status via OSC 7501 in both TUI and lean mode, allowing terminals and multiplexers to surface agent idle/working/waiting states - Exposes agent execution capabilities to callers …View on GitHub ↗ -
Release notes excerpt
This release adds GitHub repository skill loading, evaluator routing through the models gateway, and several model pricing and capability fixes, along with TUI stability improvements and expanded debug tool hook support. What's New - Adds support for loading skills from public GitHub repositories via https://github.com/owner/repo URLs in skills: entries - Runs tool_input_transform and post_tool_use hooks on debug tool calls, completing the full hook pipeline for docker agent debug tool …View on GitHub ↗ -
Release notes excerpt
This release fixes two issues: oversized tool output handling and duplicate line rendering in the lean TUI. Bug Fixes - Fixes oversized tool results from blocking chat or compaction requests by bounding complete tool results to 50 KiB and sanitizing historical tool results on outgoing requests - Fixes duplicate lines appearing while streaming wrapped text in the lean TUI by avoiding replay of archived lines when markdown changes only affect ANSI formatting --- What's Changed …View on GitHub ↗