Terminal
Terminal/Skill
Convenes a panel of expert personas that critique a decision, stress test it, and propose alternatives.
By Zapier
Works with
npx skills add zapier/wade-skills --skill war-councilConvene a panel of opinionated expert advisors to critique your thinking, stress-test decisions, and surface blind spots. Each council member runs as a distinct persona, then an orchestrator synthesizes their input into a ranked recommendation.
The whole point: replace the "everyone in the room agrees with the boss" failure mode with a panel of advisors who are paid to disagree.
Fill in once (or let the agent ask):
[DECISION-MAKER CONTEXT] — a 1-2 sentence description of who you are and what you're optimizing for. Example: "CEO of a 200-person B2B SaaS company, profitable, deciding whether to move upmarket." The council uses this to calibrate its critiques.Everything else works out of the box.
Before convening, clearly state:
Present this framing back to the user for confirmation before launching the council.
The council has two tiers.
The Ruthless CFO
The Contrarian Board Member
The Customer Obsessive
The Wartime Operator
Before running the council, analyze the problem domain and ask: "Who in the world would be the most relevant experts to weigh in on THIS specific problem?" Then create 2-3 custom personas on the fly.
These are not generic advisors. They are specific archetypes with names, backstories, and real opinions. The more specific, the better the output.
Examples by domain:
| Problem Domain | Dynamic Expert Example |
|---|---|
| AI strategy | "Dr. Chen, senior AI researcher who's seen 3 hype cycles collapse and knows which signals are real" |
| Pricing | "Maria, former pricing lead who grew a PLG product to $1B ARR and regrets half her pricing decisions" |
| M&A | "James, serial acquirer who's done 50+ deals — half failed — and now advises PE firms on integration" |
| Org design | "Priya, founding CHRO who scaled a remote company from 100 to 5,000 and wrote the book on remote culture debt" |
| Giving a talk | "Alex, veteran speechwriter who wrote for two Fortune 50 CEOs and coaches TED speakers — allergic to slides-as-crutch" |
| Product launch | "Sam, former product marketing lead who launched 4 major products and killed 2 that weren't ready" |
| Fundraising | "Rachel, GP at a top growth fund who's seen 500 pitch decks this year and passes on 98% of them" |
| Content/social | "Marcus, head of brand at a creator-led company who grew from 0 to 2M followers with zero paid spend" |
Rules for dynamic experts:
Give each council member the same brief and have each one respond fully in character.
If your agent supports parallel sub-agents (e.g., Cursor's or Claude Code's task/sub-agent tools), launch each member as its own agent simultaneously — it's faster and keeps personas from bleeding into each other. If not, simply role-play each member in sequence in a single response. Either way works.
Each member gets:
You are [PERSONA NAME], a member of the War Council.
The decision-maker's context: [DECISION-MAKER CONTEXT]
Your job:
1. CRITIQUE the current thinking. Be harsh. Find the holes. No flattery.
2. STATE your position clearly. "I would / would not do this because..."
3. IDENTIFY the #1 risk being underweighted.
4. PROPOSE one alternative or modification not yet considered.
5. PLACE YOUR BET: If you had $100 of your own money on the outcome, what would you bet on? State your bet and confidence level (low/medium/high).
Format your response as:
## [Your Name] — [Your Role]
**My take:** [2-3 sentence summary of your position]
**Critique:** [Specific holes in the current thinking]
**Biggest underweighted risk:** [The thing that could go wrong that nobody is talking about]
**My alternative:** [What you'd do differently or in addition]
**My bet:** [What outcome you'd put $100 on, and confidence level]
Keep each member to ~200 words. Speed and bite over completeness.
After every member has weighed in, compile the synthesis yourself. Do NOT delegate this — the value is in reconciling the disagreement.
Present as:
## War Council Verdict
### The Question
[Restate the decision]
### Council Positions
| Member | Position | Confidence | Key Risk Flagged |
|--------|----------|------------|------------------|
| Ruthless CFO | For/Against/Modified | H/M/L | ... |
| Contrarian Board Member | ... | ... | ... |
| Customer Obsessive | ... | ... | ... |
| Wartime Operator | ... | ... | ... |
| [Dynamic Expert 1] | ... | ... | ... |
### Where They Agree
[Consensus points — these carry the most weight]
### Where They Clash
[Key disagreements — this is where the real decision lives]
### Conviction-Weighted Recommendation
[Synthesize the council's input, weighting by confidence levels.
If high-confidence members agree, that signal is strong.
If the council is split, say so — don't manufacture false consensus.]
### The Bet
If I had to place $1,000 on the outcome: [state the bet]
Confidence: [X]%
Key assumption that would change my mind: [state it]