| name | thinking-council |
| description | Structured deliberation with 4 advisors illuminating a question from
different perspectives. Inspired by Daniel Miessler's Council pattern.
Visible transcript of the debate.
[WHAT] Convene a council of 4 distinct perspectives (expert roles fitted
to the question) that debate, challenge, and build on each other's
arguments. Produces a visible transcript with conclusion and action recommendation.
[WHEN] Use when: council, deliberate, multiple perspectives, weigh arguments,
structured debate, "I need to think this through"
[EXPERTISE] Structured deliberation, multi-perspective analysis, dialectical reasoning
|
| argument-hint | [question to deliberate] |
| allowed-tools | Read, Grep, Glob, Write, WebSearch, WebFetch |
The Council
Role: structured deliberation with 4 perspectives.
Tone: respectful but honest. Each advisor is genuine in their perspective.
How the Council works
- Question is presented (from $ARGUMENTS or the user's prompt)
- 4 advisors are selected based on the question's domain
- 3 deliberation rounds with a visible transcript
- Conclusion with consensus, dissent, and a recommendation
Selecting advisors
Pick 4 roles that MAXIMISE perspective spread. At least one must be contrarian.
Example role pools:
| Question domain | Possible advisors |
|---|
| Policy/society | Political scientist, practitioner, activist, economist |
| Tech/architecture | Architect, security expert, end user, ops engineer |
| Research | Supervisor, critical opponent, industry person, peer doctoral student |
| Strategy | Optimist, pessimist, historian, pragmatist |
| Security/threats | Defender, attacker, analyst, decision-maker |
Advisors must NOT be generic. Name them, give them background, make them specific.
Deliberation protocol
Round 1: opening statements
Each advisor presents their position in 3-5 sentences. Clear stance.
Round 2: challenge and depth
Each advisor reacts to the others' positions. Identifies blind spots, builds on, challenges assumptions. The contrarian role MUST challenge consensus.
Round 3: synthesis and closing words
Each advisor offers their closing word in light of what they've heard. Have they changed their mind? What remains?
Output format
## Council on: [The question]
### Advisors
- **[Name]** — [Role/background, 1 sentence]
- **[Name]** — [Role/background, 1 sentence]
- **[Name]** — [Role/background, 1 sentence]
- **[Name]** — [Role/background, 1 sentence]
### Round 1: opening statements
**[Name]:** [Position]
**[Name]:** [Position]
**[Name]:** [Position]
**[Name]:** [Position]
### Round 2: challenge
[Free format, advisors interact]
### Round 3: closing words
[Each advisor summarises]
### The Council's conclusion
**Consensus:** [What the council agrees on]
**Dissent:** [Where disagreement remains]
**Recommendation:** [Action recommendation to the user]
Round-by-round transcript
A visible transcript is mandatory. The Council's value lies in the debate, not the conclusion. The user must be able to follow each advisor's position round by round and see WHERE views shift or harden.
- Mark ROUND 1, ROUND 2, ROUND 3 clearly as headings
- Per round: all 4 advisors in the same order, so position shifts are visible
- In round 3: explicitly note if an advisor has changed position (and why) or hardened (why they remain unconvinced)
- The conclusion comes AFTER the transcript, not instead of it
Rules
- Advisors are NEVER unanimous from the start. If everyone agrees, swap one out.
- The contrarian role is mandatory and must be genuinely challenging, not a straw man.
- The transcript should be readable and interesting, not mechanical.
- The user can interrupt between rounds with questions or adjustments.
- Advisors should reference real thinkers, data, and examples where relevant.
