What is it?
What it is
A Claude Code skill that guides technical founders through building and deploying managed agents on Claude Managed Agents (CMA) for purposes like internal tools, product components, or customer-facing agents.
Why it exists
To educate users on composing CMA primitives step-by-step, serving as a reference implementation for creating agents with interview, staging, grading, and scheduling workflows.
Who should use it
Teams building with HTML who want an open-source, self-hosted option.
Who should avoid it
Teams that need a fully managed SaaS with enterprise SLAs out of the box.
How it works
A quick walkthrough in plain English
How launch-your-agent works
Step 1 of 3
You interact with it
Open launch-your-agent, send a request, or connect it to your stack.
Features
Advantages
- Educational: explains each step of building a Claude Managed Agent
- Helps technical founders turn ideas into live agents quickly
- Produces production-ready artifacts (API payloads, launch script, eval)
- Low cost: runs cost only cents per execution
- Open source under Apache-2.0 license
- No installation needed beyond cloning repo and using Claude Code
- Wrap-up skill provides status check and next steps
Disadvantages
- Reference implementation only; not maintained and not accepting contributions
- Token-intensive due to detailed explanations
- Requires Claude Code and an Anthropic API key (user's own account)
- Limited to HTML/JavaScript skill context; not a standalone application
- May be overkill for simple agents due to exhaustive process
Installation
native
git clone https://github.com/anthropics/launch-your-agent.git cd launch-your-agent claude
FAQ
What is the launch-your-agent skill and what does it help me build?
launch-your-agent is a Claude Code skill that guides a technical founder through building a Claude Managed Agent (CMA) – whether an internal worker, product component, or customer‑facing agent. It interviews you, scopes a v0, launches it in your own account, grades it against your definition of done, iterates, and can schedule recurring deployments, while also laying out explicit v1/v2 plans.
How do I get started with launch-your-agent?
Clone the repository, navigate into the folder, and run Claude Code: ```bash git clone https://github.com/anthropics/launch-your-agent.git cd launch-your-agent claude ``` Then type `/launch-your-agent` to start the skill. The skills in `.claude/skills/` are auto‑loaded when you run Claude Code inside this directory.
What prerequisites do I need to run the skill?
You need Claude Code installed and signed in, plus an Anthropic API key for your own account (created at platform.claude.com → API keys). The key is stored locally in a `.env` file and never sent to the chat. Running the agent incurs only a few cents of cost.
What do I walk away with after completing the flow?
You get a live managed agent in your Anthropic Console (agent, environment, graded run, and optional scheduled deployment). Additionally, a `my-agent/` folder is created containing the build sheet, exact API payloads, a resumable launch script, an eval scaffold, an overview page, and a NEXT‑DIRECTIONS.md that outlines v1/v2 upgrades.
What is the purpose of the companion `/wrap-up` command?
Running `/wrap-up` after you finish (or at any later time) refreshes the overview page, recaps every CMA primitive you now own, and suggests the next 1–2 upgrades for your agent.
Where can I find the reference documentation for Claude Managed Agents?
The definitive source for CMA API details and primitives is the official documentation at https://platform.claude.com/docs/en/managed-agents/overview. The repo also includes `cma-primitives.md` as an inventory of those primitives and their limits.
Featured in Videos
YouTube tutorials and walkthroughs for launch-your-agent
t5
70,223,459 views
Starts at 00:00
Alternatives
Similar projects ranked by category, topics, and text overlap.
