💰 The Council

A board of 7 AI advisors with one obsession: does this make you money?

Treasurer · Salesperson · Builder · Skeptic · Closer · Growth · Critic — plus a Chair who cuts through to one next action, with a date and kill-criteria.

Works in Claude Code · Kiro · Cursor · GitHub Copilot · any AGENTS.md tool.

---

Why

Makers drown in "productive procrastination" — another refactor, another feature, another tool — while the project makes $0. The Council exists to name that trap out loud and drag every decision back to one question: does this bring the first dollar closer, or is it technical entertainment disguised as progress?

Seven advisors, each answering only from their lens, plus a Chair who won't average them. You get a verdict, the shortest path to real revenue, and one action — never a comforting to-do list.

What makes it different

• Revenue-first, brutally honest. If it won't sell, the Council says so. "Built" is not "profitable," and it won't let you confuse the two.
• The Skeptic + the Critic. One asks what you're avoiding; the other refuses to accept "almost done." Together they kill vapor before you ship it.
• Vertical + lateral thinking. Every advisor does a deep pass and one lateral pass — Inversion, Pre-mortem, Second-order, Analogy, or Devil's advocate.
• The L-move. Before the verdict the Chair attacks the premise every advisor shared: "we all assumed X — what if the real business is Y?" — then discards three plays nobody proposed on the spot. The survivor is your pivot.
• One action, with kill-criteria. Not ten. The one that unblocks the nearest dollar — plus the condition under which you drop the whole thing. Without kill-criteria, every project becomes a hobby.

Install

Claude Code — one command (recommended)

/plugin marketplace add rogercg/council
/plugin install council@council

No clone, no scripts. Then use /council in any project.

Every other tool

git clone https://github.com/rogercg/council.git && cd council

./install.sh kiro      /path/to/your/repo   # Kiro     → #council
./install.sh cursor    /path/to/your/repo   # Cursor   → @council
./install.sh copilot   /path/to/your/repo   # Copilot  → /council
./install.sh agents-md /path/to/your/repo   # Codex/Gemini/Zed via AGENTS.md

Windows: powershell -ExecutionPolicy Bypass -File install.ps1 <tool> <path>.

Use

/council  Should I build a paid tier for my side project or keep it free to grow the audience?

## Council verdict: PIVOT
Agreement: … · Divergences (risk lives here): …
The L-move: … · Lateral alternative: …
Honest read: the truth you need, not the one you want.
The one next action: … by <date> · Kill criteria: …

See examples/ for a full run.

The 7 lenses

Lens	Asks
💵 Treasurer	Cost, runway, unit economics, opportunity cost
🛒 Salesperson	Who exactly pays? Is there demand and a channel?
🔨 Builder	What's the smallest MVP that can charge today?
🧨 Skeptic	Why does this fail? What are you avoiding?
🎯 Closer	Shortest path to the first $1, then the first $1,000 recurring
📣 Growth	The hook, the channels, the launch, retention
🎨 Critic	Is it actually good, or is it vapor?
⚖️ Chair	Synthesizes without averaging; one action + kill-criteria

How it's built

core/council.md   ← the canonical prompt (single source of truth)
adapters/*.md     ← per-tool frontmatter only
install.sh / .ps1 ← assemble adapter + core into your repo

Edit the Council = edit core/ and re-run install. Every tool stays in sync.

License

MIT.

---

_{A companion project: The IT Council — same board format, aimed at whether a technical decision is sound.}