⚔️ War Room

AI that argues back. Ships better.

Multi-agent sessions with a built-in devil's advocate. Specialist agents collaborate in waves, a CHAOS agent attacks every decision, and 19 structured protocols force better thinking.

Free. Open source. MIT.

---

What Happened

We ran the same project through a standard multi-agent session, then through War Room.

Same model. Same input. Different operating system.

---

Quick Start

# 1. Initialize
bash skills/war-room/scripts/init_war_room.sh my-project

# 2. Write your brief
# Edit war-rooms/my-project/BRIEF.md — describe what you're building

# 3. Inject the DNA
# Copy skills/war-room/references/dna-template.md → war-rooms/my-project/DNA.md

# 4. Run it
# Tell your agent: "Run a war room on my-project"

The agent reads the skill, picks the right specialists, runs them in waves, unleashes CHAOS after each wave, and consolidates everything into a blueprint.

---

How It Works

Agents

You pick 4-13 specialists based on your problem:

Custom roles welcome: AI-ENG, AUDIO, DATA, OPS — whatever the problem needs.

Waves

Agents don't all run at once. They run in dependency order:

Wave 1: Foundation (ARCH + SEC + PM)     → decisions that others depend on
Wave 2: Specialists (UX + AUDIO + AI)    → build on Wave 1 decisions
Wave 3: Builders (DEV + OPS)             → implement based on Wave 1+2
Wave 4: Validators (QA + MKT + CHAOS)    → stress-test everything

CHAOS shadows every wave. Not just the end.

The CHAOS Agent

The devil's advocate. Attacks assumptions. Rates decisions:

• SURVIVES — withstands scrutiny
• WOUNDED — valid but has weaknesses
• KILLED — doesn't hold up, needs rethinking

CHAOS also produces counter-proposals — alternative approaches nobody considered.

---

The Protocols

19 structured decision protocols across 4 pillars. Not suggestions — constraints that every agent must follow.

Essential 7 (start here)

Advanced 12 (power users)

The full DNA adds: Dialectic Obligation, Mirror Test, Ripple Analysis, Tension Map, Causal Chain Verification, Tempo Tagging, Create-Then-Constrain, Barbell Strategy, and Lessons Permanent.

Full protocol reference: dna-template.md

---

What It Produces

war-rooms/my-project/
├── BRIEF.md              ← Your project description
├── DNA.md                ← The operating protocols
├── DECISIONS.md          ← Append-only decision log
├── STATUS.md             ← Agent completion tracking
├── BLOCKERS.md           ← Issues requiring human input
├── TLDR.md               ← Executive summary
├── agents/
│   ├── arch/             ← Architecture specs
│   ├── pm/               ← Product requirements
│   ├── chaos/            ← Challenges + counter-proposals
│   └── [role]/           ← Any specialist
├── artifacts/
│   └── BLUEPRINT.md      ← Consolidated output
├── comms/                ← Inter-agent messages
└── lessons/              ← Post-mortem learnings

---

When To Use It

Use it when:

• Decisions cost weeks of work if wrong
• You need multiple perspectives but don't have multiple people
• You need a PRD, architecture, or strategy that survives contact with reality
• You want to stress-test an existing plan before committing

Don't use it when:

• The task is simple and well-defined
• You need a quick answer, not deep analysis
• You've already decided and just need execution

---

License

MIT. Use it, fork it, build on it.

---

"The unexamined life is not worth living." — Socrates

"Wind extinguishes a candle and energizes fire." — Nassim Taleb

"O melhor conhecimento é aquele que é passado adiante." — Max Kleinz