@eventmodelers/agent-modeling-kit

Real-time Claude agent + skill kit for the Eventmodelers platform.

Quick start

npx @eventmodelers/agent-modeling-kit install

The installer will prompt for your API token and Organization ID from app.eventmodelers.ai/account.

What gets installed

your-project/
├── .agent-modeling-kit/          ← all agent files (created by install)
│   ├── .eventmodelers/
│   │   └── config.json           ← your token + org (gitignored)
│   ├── realtime-agent/           ← Node.js agent (Supabase Realtime)
│   ├── ralph.sh                  ← agent loop
│   └── agent.sh                  ← claude runner
├── .claude/
│   └── skills/                   ← eventmodelers skills for Claude Code
└── CLAUDE.md                     ← agent instructions

Claude execution & config resolution

Custom Anthropic endpoint / model

During install you can optionally point the agent at a local LLM server (vLLM, Ollama) instead of the default Claude Code endpoint, and/or pin a specific model. The installer shows an arrow-key select:

🧠 Configuring Claude execution (optional)...
  ● None — use the default Claude Code endpoint
  ○ Local vLLM   (http://localhost:8000)
  ○ Local Ollama (http://localhost:11434)
  ○ Custom…

Pick an option, then optionally enter a model (e.g. claude-sonnet-5). Both are stored alongside your credentials in .agent-modeling-kit/.eventmodelers/config.json:

{
  "organizationId": "...",
  "token": "...",
  "anthropicBaseUrl": "http://localhost:8000",
  "model": "claude-sonnet-5"
}

You can edit this file by hand at any time — ralph-claude.js re-reads it on every run. When anthropicBaseUrl is set it's exported as ANTHROPIC_BASE_URL for the claude process; when model is set it's passed as claude --model <model>. Omit or delete either field to fall back to the default Claude Code setup.

Hierarchical config resolution

.eventmodelers/config.json is resolved by walking from .agent-modeling-kit/ up through every parent directory, merging fields as it goes:

• A value set by a closer (more specific) directory always wins over one set farther up.
• The walk stops as soon as token, organizationId, and baseUrl are all resolved — it won't keep climbing just to find anthropicBaseUrl/model.
• anthropicBaseUrl/model are picked up along the way if a directory the walk passes through happens to set them, but they never force the walk to continue further up.

This lets you keep a shared base config higher up the tree (org token, anthropicBaseUrl, model) while each project only overrides what's specific to it:

~/workspace/.eventmodelers/config.json                              ← shared: token, organizationId, anthropicBaseUrl, model
~/workspace/project-a/.agent-modeling-kit/.eventmodelers/config.json  ← project-specific overrides, if any

If a project's own config already has everything needed to connect, the walk stops there and the shared config higher up is never read.

Running the agent

Run both in separate terminals from your project root:

# Terminal 1 — realtime agent (listens for prompts → writes tasks.json)
cd .agent-modeling-kit/realtime-agent && npm run dev

# Terminal 2 — agent loop (reads tasks.json → runs claude in project root)
cd .agent-modeling-kit && ./ralph.sh

The loop skips when tasks.json is empty. Claude always runs in your project root so it can access your full codebase.

Skills

Use skills in Claude Code with /skill-name:

| Skill | Description | |-------|-------------| | /connect | Set up board connection | | /timeline | Live event storming facilitator | | /wdyt | Business analyst review of your event model | | /storyboard | Build a full visual storyboard | | /storyboard-screen | Design individual wireframe screens | | /place-element | Place commands/events/read models on the board | | /learn-eventmodelers-api | Full API reference for agent use | | /attributes | Add/rename attributes across a chain of elements | | /examples | Add example data to element fields | | /update-slice-status | Update slice status on the board |

Commands

npx @eventmodelers/agent-modeling-kit install    # install + configure
npx @eventmodelers/agent-modeling-kit status     # check what's installed
npx @eventmodelers/agent-modeling-kit uninstall  # remove installed files

Contributors

| Contributor | Contribution | |-------------|-------------| | Yordis Pietro | All eventmodeling-* skills |