feat: initial agent-builder plugin (v0.1.0)

Build complete autonomous agent systems with Claude Code.
7-phase guided workflow: map work, CLAUDE.md, agent team,
pipeline, security, deployment, test.

Components:
- commands/build.md: main guided workflow
- agents/builder.md: scaffolding agent
- skills/agent-system-design: architecture knowledge + 4 references
- scripts/templates: hooks, automation, launchd, systemd

Covers 22 OpenClaw capabilities across 4 deployment targets
(local, Mac Mini, VPS, Managed Agents).

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
Kjell Tore Guttormsen 2026-04-10 19:10:54 +02:00
commit 075383990f
17 changed files with 1895 additions and 0 deletions

67
README.md Normal file
View file

@ -0,0 +1,67 @@
# Agent Builder
Build complete autonomous agent systems with Claude Code.
OpenClaw gives you a pre-built agent. This plugin helps you
build your own, tailored to your work and deployed where you
choose.
## Install
```
/install agent-builder
```
Or for local development:
```
claude --plugin-dir /path/to/agent-builder
```
## Commands
| Command | What it does |
|---------|-------------|
| `/agent-builder:build` | Guided 7-phase workflow to build a complete agent system |
| `/agent-builder:deploy` | Configure deployment for your chosen platform |
| `/agent-builder:evaluate` | Assess your setup against 22 agent capabilities |
| `/agent-builder:status` | Quick status check of your agent infrastructure |
## What `/agent-builder:build` creates
The build command walks you through 7 phases:
1. **Map your work** -- identify recurring tasks and agent roles
2. **Operating manual** -- generate a CLAUDE.md tailored to your domain
3. **Agent team** -- scaffold specialized agents in `.claude/agents/`
4. **Pipeline** -- create skills that orchestrate your agents
5. **Security** -- add hooks for command blocking and audit logging
6. **Deployment** -- configure for your platform (local, Mac Mini, VPS, or Managed Agents)
7. **Test** -- run on real work and iterate
Each phase produces files in your project. You can stop after any
phase and resume later.
## Deployment targets
| Platform | Best for | Always-on |
|----------|---------|-----------|
| Local (cron/launchd) | Development, testing | With tmux |
| Mac Mini (launchd + channels) | Personal agent | Yes |
| VPS (systemd + cron) | Server agent, teams | Yes |
| Managed Agents (Anthropic API) | Production, scaling | Yes (cloud) |
## What this covers
Maps all 22 OpenClaw capabilities to Claude Code equivalents:
13 full matches, 8 different approaches, 1 gap (Canvas/A2UI).
Run `/agent-builder:evaluate` to see your current coverage.
## Requirements
- Claude Code v2.1.86 or later (for full feature coverage)
- The `plugin-dev` plugin (recommended for agent/skill creation)
## License
MIT