Cadre — AI skill for Claude Code
Coordinated Agent Deployment Runtime Engine.
How to install Cadre
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open ChristopherKahler/cadre and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Cadre does
Coordinated Agent Deployment Runtime Engine. A framework for running a company of AI agents with real org structure. members, roles, contracts, operations, goals, board-gated decisions, atomic work units. Solo-operator scoped, runtime-agnostic. Claude Code native, designed to swap.
Alternatives in AI
- Paca — AI-native, free, open-source alternative to Jira, Trello, ClickUp & Monday 1.8k ★
- Adopt AI Properly — Workflow recipe — the org-side AI adoption arc, policy to proof, by chaining 4 skills 1.3k ★
- Agency Orchestrator — 🚀 One sentence → multi-AI-role collaboration → complete plan in minutes 453 ★
README
Cadre
*Coordinated Agent Deployment Runtime Engine*
**Run your AI agents like a company.**
Cadre is a framework for orchestrating a persistent roster of AI Members with roles, ownership, atomic work assignment, and board-level approval gates. Built for Claude Code, designed runtime-agnostic from v1 via a formal Contract interface (swap to OpenClaw, Codex, or any agent runtime that can execute a prompt).
pip install cadre # (once published to PyPI; dev install below)
cadre init . --demo --install-hooks
Open the directory in Claude Code. The session-start hook injects your active roster, pending approvals, and goal health into every session.
Why this exists
Every Claude Code session starts cold. You re-explain context, re-assign work, re-approve the same decisions. The moment you try to run more than one agent or split work across sessions, the whole thing falls apart: who owns what? which Unit is claimed? who's overloaded? who decides when to hire?
Cadre gives you a persistent structure that survives between sessions — a Firm of Members, each with a Contract (runtime), claiming Units (atomic work) under Operations (departments) toward Goals (KPIs), with Gates (your approval) for significant actions. The framework tracks it. The session hook surfaces it. You stop hand-driving and start governing.
Core vocabulary
| Entity | Analogy | What it does |
|---|---|---|
| Firm | Your company | One Cadre install = one Firm. Has a name (e.g., chrisai). |
| Member | Employee | An AI agent with a role (Writer, Editor, CMO). Named (Quill, Sterling, Sage). |
| Contract | Employment agreement | How a Member executes: which runtime (Claude Code / OpenClaw / Codex), which skills/commands they can use. |
| Operation | Department | Long-running business function ("Content Publishing"). |
| Project | Initiative | Scoped deliverable inside an Operation ("Blog Cadence — First 8 Posts"). |
| Unit | Ticket | Atomic work |
Related Skills
Trupe
Your AI coding agents, performing together. Trupe turns the CLIs you already have — Claude Code, Codex, Gemini
Tavern
A Mac app for a party of AI agents: Pi agents with durable threads, hosted members over mail and iMessage, Lin
5dive
Run a company of AI agents on a server you own. Spin up named agents (claude, codex, pi…), put them on an org
Org OS Starter
Company OS as Code — 会社の運営知識・規範・意思決定をひとつのリポジトリに構造化し、人とAIエージェントが同じ基盤で働くためのファイル規約の参照テンプレート。a company brain / AI-
LLMs Txt Checker
Validates llms.txt files for AI agent discoverability. Checks structure, format, required fields, and best pra
SpielOS
An open-source AI company operating system: Goals, Runs, Workers, and AI departments under one Director loop (
Related Agents
Church Validation Engineer
P1 deterministic validation scripts: validate_members.py (M1-M6), validate_finance.py (F1-F5), validate_schedu
Case Engineer
Owns the 3D-printed enclosure: code-CAD model, fastening stack-up, air-gap ledger, printability. Built from ge
Org Designer
Head of People. Continuously evaluates team shape. Detects cracks, gaps, bloat, recurring failures. Proposes s