consciousness-council

Trigger phrases

Phrases that activate this skill when typed to Claude Code:

• council mode
• deliberate on this
• what would different experts think
• help me think through this from all sides
• devil's advocate

What it does

consciousness-council is a Claude Code skill from K-Dense AI’s scientific-agent-skills repo, originally authored by AHK Strategies. It turns Claude into a structured deliberation chamber: instead of one voice giving one answer, the skill summons 4–6 distinct thinking archetypes, forces them into genuine tension with each other, and synthesises their positions into a concrete recommendation.

The twelve archetypes include The Architect (systems thinking), The Contrarian (inversion), The Empiricist (evidence-first), The Ethicist (consequence-aware), The Futurist (second-order effects), The Pragmatist (action-oriented), The Historian (precedent), The Empath (human-centred), and The Outsider (naive cross-domain questions). The skill selects whichever 4–6 will produce the most productive tension for your specific question — agreement is cheap, useful friction is the point.

A session produces three things: each archetype’s position stated honestly (including their own blind spots), the points where they directly contradict each other, and a synthesis that names the trade-offs rather than papering over them.

When to use it

Reach for it when:

• You have a decision with no obvious right answer and can feel yourself circling the same frame
• You want structured devil’s advocate pressure before committing to a direction
• A team or stakeholder has a strong prior and you need to surface what it misses

When not to reach for it:

• Mechanical, well-scoped tasks where the path is clear — the deliberation overhead costs more than it returns
• You’ve already decided and just want execution; the Council interrogates, it doesn’t rubber-stamp

Install

Copy the consciousness-council SKILL.md into your project’s .claude/skills/consciousness-council/ directory.

Triggers on: “council mode”, “deliberate on this”, “what would different experts think”, “help me think through this from all sides”, “devil’s advocate”. Also activates when the user describes a dilemma or trade-off with no obvious answer.

What a session looks like

A typical session has three phases:

1. Summon. Based on your question, the skill selects 4–6 archetypes whose reasoning styles will genuinely clash. It names each one and briefly states their starting orientation before anyone speaks.
2. Deliberate. Each archetype states their position, including the question they would ask that no one else is asking and the blind spot they acknowledge in their own view. Where two archetypes directly contradict, the skill surfaces the exact point of disagreement rather than smoothing it over.
3. Synthesise. The skill names what the archetypes agree on, where the genuine trade-offs live, and what decision or next step the synthesis points toward — without pretending the tension has been resolved when it hasn’t.

The discipline that makes it work: archetypes are chosen to clash, not to confirm. A Council that agrees is a waste of a session.

Receipts

Where it works well:

• Strategy and positioning questions where the right answer depends heavily on which values you weight — the Ethicist and Pragmatist will surface exactly that tension
• Architecture and build-vs-buy decisions where The Architect and The Pragmatist tend to genuinely disagree in useful ways
• Any question where your instinct is “I already know the answer” — The Contrarian is most useful there

Where it backfires:

• Tactical execution questions (“how do I write this function”) — too much ceremony for too little return
• When the user is looking for validation rather than deliberation; the Council will not provide it

Pattern that works: state the question as a genuine dilemma (“should we do X or Y, given Z constraint”) rather than as a brief. The more the framing acknowledges the tension, the sharper the Council’s responses.

Source and attribution

Originally authored by AHK Strategies (ashrafkahoush-ux), distributed via K-Dense AI’s scientific-agent-skills repository. The canonical SKILL.md lives in the consciousness-council folder.

License: MIT. Install, adapt, and redistribute with attribution preserved.

This page documents the skill from a practitioner’s perspective. For the formal spec and updates, defer to the source repo.