# anthropics/launch-your-agent architecture
> Architecture diagram of the GitHub repository anthropics/launch-your-agent, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-06-20). It is an AI-made overview: check details against the source.
- Interactive diagram: https://gitdiagram.com/anthropics/launch-your-agent
- Repository: https://github.com/anthropics/launch-your-agent
## Overview
This repository is not a runtime application; it is a **Claude Code skill package** that orchestrates a human-guided build flow for **Claude Managed Agents (CMA)**. The architecture is mostly documentation, prompts, templates, and reference artifacts that Claude Code uses to drive the process end to end.
1. **Primary execution boundary: Claude Code + skills**
- The main entrypoint is the `/launch-your-agent` skill under `.claude/skills/launch-your-agent/`.
- Claude Code loads the skill automatically when run from this repo, so the repo functions as an operational playbook rather than a compiled app.
- The companion `/wrap-up` skill provides post-build closeout and status review.
2. **Main orchestration flow**
- The launch skill is organized into four phases: **interview → stage & launch → grade & iterate → run without you**.
- This creates a staged control flow where the assistant first elicits requirements, then turns them into a CMA build, then validates the result, and finally sets up unattended operation if the agent should recur.
- The flow is explicitly designed to produce a v0 plus a forward plan for v1/v2 rather than trying to solve everything in one pass.
3. **External service boundary: Claude Managed Agents**
- The real system being configured lives in **Claude Managed Agents** on the Anthropic platform.
- The repo’s skill references the CMA API docs and primitive definitions as the source of truth for what can be created, launched, graded, and scheduled.
- The actual agent, environment, graded runs, and scheduled deployment are created in the user’s own Anthropic account, not inside this repository.
4. **Interview and scoping subsystem**
- The skill’s reference materials include interview guidance and mapping logic so the conversation can be turned into a concrete build sheet.
- This is the requirements-to-architecture translation layer: user intent becomes CMA primitives, launch configuration, and success criteria.
- The repo emphasizes scoping to a versioned build plan, which is why the output includes explicit next steps instead of only a finished launch.
5. **Launch and iteration subsystem**
- The core build phase generates the artifacts needed to stage and launch the agent, then checks whether the result matches the stated definition of done.
- The iteration loop exists to improve the agent against user-defined evaluation criteria, not just to confirm that deployment succeeded.
- This means the system has a feedback loop between configuration, execution, and grading.
6. **Generated local artifact boundary**
- The README says the flow leaves behind a `my-agent/` folder containing the build sheet, exact API payloads, resumable launch script, eval scaffold, overview page, and next-direction plan.
- That folder is the durable handoff from the interactive skill to a concrete project state the user owns.
- In architecture terms, it is the boundary where the ephemeral conversation becomes persisted implementation assets.
7. **Reference knowledge base**
- `.claude/skills/launch-your-agent/references/` contains the skill’s supporting knowledge: verified API call shapes, examples bank, mock connectors, interview notes, and an overview template.
- `cma-primitives.md` at the repo root provides a compact inventory of CMA capabilities and limits.
- These files are the system’s “design-time schema” for deciding what can be built and how.
8. **UI and presentation artifacts**
- The `ui/` directory contains example assets for the generated overview page and build sheet JSON.
- `ui/agent-overview-example.html` is a presentation template for the post-build overview page.
- `ui/build-sheet.example.json` shows the structured shape of the configuration the skill produces or consumes.
- This is not the application runtime; it is the human-facing reporting layer for the built agent.
9. **Documentation and governance layer**
- `README.md`, `CLAUDE.md`, `CONTRIBUTING.md`, and `LICENSE` define usage, operating assumptions, and project boundaries.
- The README explicitly states this is a **reference implementation** and not maintained, which matters for architecture because it frames the repo as instructional scaffolding rather than an evolving product.
- The docs are part of the repo’s control plane for how the skill should be used.
10. **Operational inputs and secrets**
- The workflow requires an Anthropic API key in a local `.env` file for the user’s account.
- That secret is an external input to the launch process and must stay outside chat and out of versioned artifacts.
- This is the only material secret-management boundary called out in the repo.
11. **End-state and output model**
- Successful use of the repo yields a live CMA agent plus a local package of build artifacts and a documented upgrade path.
- The architecture therefore produces two outputs: a deployed remote agent in Anthropic’s platform and a local handoff bundle for the owner.
- The `/wrap-up` skill closes the loop by regenerating the overview page and recapping owned primitives.
12. **How to draw it**
- A clean diagram would show: **User → Claude Code → launch skill references/templates → CMA API / Anthropic platform → generated local `my-agent/` assets**, with a separate **wrap-up** path fed by the deployed agent state and local artifacts.
- The repo itself sits mostly in the orchestration and documentation layer, while the actual execution happens in the external CMA service and in the user’s local Claude Code session.
Main entry points: [`.claude/skills/launch-your-agent/SKILL.md`](https://github.com/anthropics/launch-your-agent/blob/main/.claude/skills/launch-your-agent/SKILL.md) (Launch skill), [`.claude/skills/wrap-up/SKILL.md`](https://github.com/anthropics/launch-your-agent/blob/main/.claude/skills/wrap-up/SKILL.md) (Wrap-up skill), [`.claude/skills/launch-your-agent/references/overview-template.html`](https://github.com/anthropics/launch-your-agent/blob/main/.claude/skills/launch-your-agent/references/overview-template.html) (Overview template), [`README.md`](https://github.com/anthropics/launch-your-agent/blob/main/README.md) (README)
## Diagram (Mermaid)
```mermaid
flowchart TD
subgraph group_g1["Orchestration"]
node_n2["Claude Code
skill host"]
node_n3["Launch skill
orchestrator
[SKILL.md]"]
node_n4["Wrap-up skill
closeout
[SKILL.md]"]
end
subgraph group_g2["Reference base"]
node_n5["Interview guide
requirements mapping
[interview.md]"]
node_n6["CMA API
service contract
[cma-api.md]"]
node_n7["Examples bank
reference examples
[examples-bank.md]"]
node_n8["Mock connectors
simulated integrations
[mock-connectors.md]"]
node_n9["Overview template
presentation template"]
end
subgraph group_g3["Repo docs"]
node_n10["Primitives
capability catalog
[cma-primitives.md]"]
node_n11["README
usage guide
[README.md]"]
node_n12["Claude guide
operating notes
[CLAUDE.md]"]
node_n13["Contributing
governance
[CONTRIBUTING.md]"]
end
subgraph group_g4["Presentation"]
node_n14["UI examples
sample assets"]
node_n15["Overview example
html example"]
node_n16["Build sheet example
json example"]
end
subgraph group_g5["External runtime"]
node_n17[("Anthropic CMA
external platform")]
node_n18["Local handoff
generated bundle"]
end
node_n1(("User
human operator"))
node_n1 -->|"works in"| node_n2
node_n2 -->|"loads"| node_n3
node_n3 -->|"interviews with"| node_n5
node_n3 -->|"consults"| node_n6
node_n3 -->|"patterns"| node_n7
node_n3 -->|"validates with"| node_n8
node_n3 -->|"constrains"| node_n10
node_n3 -->|"creates in"| node_n17
node_n3 -->|"writes"| node_n18
node_n3 -->|"renders"| node_n9
node_n9 -->|"influences"| node_n15
node_n3 -->|"emits"| node_n16
node_n2 -->|"runs"| node_n4
node_n4 -->|"reviews"| node_n18
node_n4 -->|"checks state"| node_n17
node_n11 -->|"describes"| node_n3
node_n12 -->|"operationalizes"| node_n3
node_n13 -.->|"governs"| node_n11
node_n14 -->|"includes"| node_n15
node_n14 -->|"includes"| node_n16
node_n3 -.->|"borrows"| node_n14
click node_n3 "https://github.com/anthropics/launch-your-agent/blob/main/.claude/skills/launch-your-agent/SKILL.md"
click node_n4 "https://github.com/anthropics/launch-your-agent/blob/main/.claude/skills/wrap-up/SKILL.md"
click node_n5 "https://github.com/anthropics/launch-your-agent/blob/main/.claude/skills/launch-your-agent/references/interview.md"
click node_n6 "https://github.com/anthropics/launch-your-agent/blob/main/.claude/skills/launch-your-agent/references/cma-api.md"
click node_n7 "https://github.com/anthropics/launch-your-agent/blob/main/.claude/skills/launch-your-agent/references/examples-bank.md"
click node_n8 "https://github.com/anthropics/launch-your-agent/blob/main/.claude/skills/launch-your-agent/references/mock-connectors.md"
click node_n9 "https://github.com/anthropics/launch-your-agent/blob/main/.claude/skills/launch-your-agent/references/overview-template.html"
click node_n10 "https://github.com/anthropics/launch-your-agent/blob/main/cma-primitives.md"
click node_n11 "https://github.com/anthropics/launch-your-agent/blob/main/README.md"
click node_n12 "https://github.com/anthropics/launch-your-agent/blob/main/CLAUDE.md"
click node_n13 "https://github.com/anthropics/launch-your-agent/blob/main/CONTRIBUTING.md"
click node_n15 "https://github.com/anthropics/launch-your-agent/blob/main/ui/agent-overview-example.html"
click node_n16 "https://github.com/anthropics/launch-your-agent/blob/main/ui/build-sheet.example.json"
classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_n2,node_n3,node_n4 toneBlue
class node_n5,node_n6,node_n7,node_n8,node_n9 toneAmber
class node_n10,node_n11,node_n12,node_n13 toneMint
class node_n14,node_n15,node_n16 toneRose
class node_n17,node_n18 toneIndigo
class node_n1 toneNeutral
```
## Components
### Orchestration
Interactive skill flow that turns a conversation into a build and wrap-up process.
- **Claude Code**: skill host · Local assistant session that loads and executes the repo skills.
- [Launch skill](https://github.com/anthropics/launch-your-agent/blob/main/.claude/skills/launch-your-agent/SKILL.md): orchestrator · Main /launch-your-agent skill that runs interview, launch, grading, and unattended-operation phases. · `.claude/skills/launch-your-agent/SKILL.md`
- [Wrap-up skill](https://github.com/anthropics/launch-your-agent/blob/main/.claude/skills/wrap-up/SKILL.md): closeout · Post-build skill that reviews status and refreshes the handoff artifacts. · `.claude/skills/wrap-up/SKILL.md`
### Reference base
Design-time materials the skills consult to map intent to CMA primitives and API calls.
- [Interview guide](https://github.com/anthropics/launch-your-agent/blob/main/.claude/skills/launch-your-agent/references/interview.md): requirements mapping · Prompts and guidance for turning user intent into a concrete build sheet. · `.claude/skills/launch-your-agent/references/interview.md`
- [CMA API](https://github.com/anthropics/launch-your-agent/blob/main/.claude/skills/launch-your-agent/references/cma-api.md): service contract · Reference for the callable API shapes and launch primitives used by the build flow. · `.claude/skills/launch-your-agent/references/cma-api.md`
- [Examples bank](https://github.com/anthropics/launch-your-agent/blob/main/.claude/skills/launch-your-agent/references/examples-bank.md): reference examples · Worked examples that show how the skill should structure outputs and decisions. · `.claude/skills/launch-your-agent/references/examples-bank.md`
- [Mock connectors](https://github.com/anthropics/launch-your-agent/blob/main/.claude/skills/launch-your-agent/references/mock-connectors.md): simulated integrations · Test doubles and connector patterns used to reason about or validate integrations. · `.claude/skills/launch-your-agent/references/mock-connectors.md`
- [Overview template](https://github.com/anthropics/launch-your-agent/blob/main/.claude/skills/launch-your-agent/references/overview-template.html): presentation template · HTML template for the post-build overview page. · `.claude/skills/launch-your-agent/references/overview-template.html`
### Repo docs
Project-level guidance and operating assumptions for the package.
- [Primitives](https://github.com/anthropics/launch-your-agent/blob/main/cma-primitives.md): capability catalog · Compact inventory of CMA capabilities and limits used as the design-time schema. · `cma-primitives.md`
- [README](https://github.com/anthropics/launch-your-agent/blob/main/README.md): usage guide · Top-level description of purpose, workflow, and repository boundaries. · `README.md`
- [Claude guide](https://github.com/anthropics/launch-your-agent/blob/main/CLAUDE.md): operating notes · Repo-specific instructions for using Claude Code with this package. · `CLAUDE.md`
- [Contributing](https://github.com/anthropics/launch-your-agent/blob/main/CONTRIBUTING.md): governance · Contribution and maintenance expectations for the reference implementation. · `CONTRIBUTING.md`
### Presentation
Example artifacts for the generated overview and build sheet.
- **UI examples**: sample assets · Example generated outputs for the overview page and build-sheet structure.
- [Overview example](https://github.com/anthropics/launch-your-agent/blob/main/ui/agent-overview-example.html): html example · Example rendered agent overview page used as a presentation reference. · `ui/agent-overview-example.html`
- [Build sheet example](https://github.com/anthropics/launch-your-agent/blob/main/ui/build-sheet.example.json): json example · Example structured build-sheet artifact for configuration and handoff. · `ui/build-sheet.example.json`
### External runtime
Where the actual Claude Managed Agent is created and run.
- **Anthropic CMA**: external platform · Remote Claude Managed Agents service where the actual agent is created and run.
- **Local handoff**: generated bundle · Persisted my-agent/ artifact set created by the launch flow and owned by the user.
### Other components
- **User**: human operator · Provides goals, requirements, and feedback for the agent build.
## Connections
- User → Claude Code: works in · The user interacts with the local Claude Code session.
- Claude Code → Launch skill: loads · Claude Code auto-loads the main launch skill from the repo.
- Launch skill → Interview guide: interviews with · The launch skill uses the interview guide to scope the build.
- Launch skill → CMA API: consults · The skill maps intent to CMA primitives and API calls.
- Launch skill → Examples bank: patterns · Examples inform expected build outputs and decisions.
- Launch skill → Mock connectors: validates with · Mock connectors support integration reasoning and checks.
- Launch skill → Primitives: constrains · The primitive catalog defines what can be built and launched.
- Launch skill → Anthropic CMA: creates in · Launch and grading actions target the remote CMA platform.
- Launch skill → Local handoff: writes · The flow leaves behind a local my-agent handoff bundle.
- Launch skill → Overview template: renders · The overview template supports the post-build summary page.
- Overview template → Overview example: influences · The template shapes the generated overview page example.
- Launch skill → Build sheet example: emits · The skill produces or consumes structured build-sheet JSON.
- Claude Code → Wrap-up skill: runs · Claude Code can invoke the wrap-up skill after build completion.
- Wrap-up skill → Local handoff: reviews · Wrap-up closes the loop against the persisted handoff bundle.
- Wrap-up skill → Anthropic CMA: checks state · Wrap-up can recap deployed agent status from the external platform.
- README → Launch skill: describes · The README documents the launch workflow and intent.
- Claude guide → Launch skill: operationalizes · Claude-specific guidance anchors the skill usage model.
- Contributing → README: governs · Contribution guidance frames the repo as a reference implementation.
- UI examples → Overview example: includes · The UI examples directory provides the overview page sample.
- UI examples → Build sheet example: includes · The UI examples directory provides the build-sheet sample.
- Launch skill → UI examples: borrows · The launch flow uses UI examples as presentation references.