[opensourcedex]

council-of-high-intelligence is a free, open source ai development platforms project written in Shell and released under MIT. It has 4,740 GitHub stars, 447 forks and 16 open issues, and was last pushed 17 hours ago. On this registry it ranks #109 of 155 tracked projects in AI Development Platforms, with 5 head-to-head comparisons available. It gained 208 stars over the last 30 tracked days.

What is council-of-high-intelligence?

Council of High Intelligence is an MIT-licensed, Shell-based agent skill for structured multi-perspective deliberation that runs full councils, focused triads, or duo debates across Claude Code, Codex, Gemini CLI, and OpenCode, built for anyone facing a hard decision that deserves more than one reasoning path.

What it is

Council of High Intelligence lives in the ecosystem of AI coding agents and multi-LLM clients: Claude Code, Codex, Gemini CLI, and OpenCode. The project assigns a hard question to deliberately different analytical personas, keeps their first positions independent, forces direct disagreement, and returns a verdict that preserves unresolved questions, dissent, kill criteria, and the next concrete action. It is an agent skill and deliberation protocol rather than a model or a hosted service, and its canonical routing table lives in SKILL.md. The command remains /council on every supported host.

The concrete problem it solves is the single reasoning path. For decisions with material downside, competing values, incomplete evidence, or irreversible consequences, a direct answer or a long undifferentiated chat can hide disagreement and manufacture false consensus. Council of High Intelligence replaces that default with a fixed round budget: members restate the problem, analyze blind, cross-examine other positions, declare a final stance, and enter synthesis. Enforcement checks look for premature agreement, repeated claims, missing dissent, and unsupported confidence. Verdicts lead with what remains unresolved, and when the weighted tally remains split, the result returns that split instead of turning it into prose consensus.

Key capabilities

  • Run /council for the full council, /council --quick for a faster panel, or /council --duo for a two-member dialectic on Claude Code, Codex, Gemini CLI, or OpenCode.
  • Select named triads with --triad, including strategy, risk, shipping, and ai-product, from domains such as architecture, ethics, debugging, product, founder, ai, ai-safety, decision, systems, uncertainty, design, economics, and bias.
  • Choose specific participants with --members torvalds,ada for a focused duo or panel.
  • Follow a fixed protocol of independent analysis, cross-examination, final stance, and synthesis.
  • Rely on enforcement checks that detect premature agreement, repeated claims, missing dissent, and unsupported confidence.
  • Receive verdicts that preserve unresolved questions, dissent, acceptable compromises, kill criteria, and one concrete next step.
  • Separate evidence with FACT, INFERENCE, ASSUMPTION, and UNKNOWN labels.

Who uses it and how

  • Claude Code users install the plugin from the marketplace and run /council inside the client.
  • Codex, Gemini CLI, and OpenCode users clone the repository once, run ./install.sh --codex-only, ./install.sh --gemini-only, or ./install.sh --opencode-only, then restart the target client.
  • Teams facing high-stakes, irreversible, or competing-frame decisions use --full so competing frames make contact.
  • Users who need breadth without a full adversarial round use --quick; when one polarity defines the decision, they use --duo.
  • Domain-focused users select three relevant lenses with --triad strategy, --triad risk, or --triad ai-product without assembling the whole council.

Getting started

Claude Code users can install directly with /plugin marketplace add 0xNyk/council-of-high-intelligence and /plugin install council@council-of-high-intelligence. For Codex, Gemini CLI, or OpenCode, clone https://github.com/0xNyk/council-of-high-intelligence.git and run ./install.sh --codex-only, ./install.sh --gemini-only, or ./install.sh --opencode-only, then restart the client.

How it compares

The provided facts do not list paid products that this project replaces, and they name no similar deliberation tools. Council of High Intelligence therefore stands alone in this registry, with its supported hosts being Claude Code, Codex, Gemini CLI, and OpenCode rather than competing products.

When to use it — and when not to

Use it when the decision has material downside, competing values, incomplete evidence, or irreversibility, and write down the decision, constraints, evidence, reversibility, and deadline before adding personas. Do not use it for factual lookups, cheap reversible experiments, or to manufacture support for a preferred answer. Self-hosting requires cloning the repository and running the installer for the chosen client; no database, storage, or SMTP dependency is mentioned, and the provided facts describe no hosted service or release history beyond the last push on 2026-09-14, so teams wanting a managed option or long release record should look elsewhere.

project readme (upstream, from github) — read inline

Council of High Intelligence

Structured multi-perspective deliberation for decisions that deserve more than one reasoning path.

Lint Release License

Claude Code · Codex · Gemini CLI · OpenCode

The council assigns a hard question to deliberately different analytical personas, keeps their first positions independent, forces direct disagreement, and returns a verdict that preserves unresolved questions, dissent, kill criteria, and the next concrete action.

Try a council

Claude Code users can install the plugin directly:

/plugin marketplace add 0xNyk/council-of-high-intelligence
/plugin install council@council-of-high-intelligence

Then convene the full council, a faster panel, or a two-member dialectic:

/council Should we open-source our agent framework?
/council --quick Should we add caching here?
/council --duo Should we use microservices or a monolith?

For Codex, Gemini CLI, or OpenCode, clone once and use the matching installer flag:

git clone https://github.com/0xNyk/council-of-high-intelligence.git
cd council-of-high-intelligence

./install.sh --codex-only
./install.sh --gemini-only
./install.sh --opencode-only

Restart the target client after installation. The command remains /council on every supported host.

Use it for the decision boundary

Good council questions have material downside, competing values, incomplete evidence, or an irreversible choice. Write down the decision, constraints, evidence, reversibility, and deadline before adding personas.

Decision field notes

The field notes distinguish facts from inference, assumptions, and unknowns. They also make it harder to use a long deliberation as decoration for a decision already made.

When to use something smaller

  • Use a direct answer or primary documentation for factual lookups.
  • Run an experiment when the choice is cheap and reversible.
  • Use --quick when cross-examination will not change the outcome.
  • Use --duo when one tension matters more than broad coverage.
  • Do not convene a council to manufacture support for a preferred answer.

Choose a mode

Mode Shape Use it when
Full Independent analysis, cross-examination, final stance, synthesis Stakes are high and competing frames need contact
Quick Restate, rapid analysis, final positions The decision needs breadth but not a full adversarial round
Duo Opening positions, direct response, final statements One polarity defines the decision

Council mode selector

/council --full What is the right pricing model?
/council --quick --triad shipping Should we release today?
/council --duo --members torvalds,ada Is this abstraction worth it?

Named triads select three relevant lenses without assembling the whole council:

/council --triad strategy Where is our defensible advantage?
/council --triad risk What could make this launch irreversible?
/council --triad ai-product Which capability belongs in the product?

Available domains include architecture, strategy, ethics, debugging, risk, shipping, product, founder, ai, ai-product, ai-safety, decision, systems, uncertainty, design, economics, and bias. The canonical routing table lives in SKILL.md.

What the protocol protects

The full protocol has a fixed round budget. Members first restate the problem, analyze blind, cross-examine other positions, declare a final stance, and then enter synthesis. Enforcement checks look for premature agreement, repeated claims, missing dissent, and unsupported confidence.

Verdict blueprint

Verdicts lead with what remains unresolved. A recommendation is paired with acceptable compromises, kill criteria, and one concrete next step. When the weighted tally remains split, the result returns that split to the user instead of turning it into prose consensus.

Evidence labels separate:

  • FACT: directly supported by supplied or retrieved evidence;
  • INFERENCE: follows from evidence but is not directly observed;
  • ASSUMPTION: required for the argument and still unverified;
  • UNKNOWN: missing information that could change the decision.

Track the outcome

A verdict is useful only if it can be checked later. Before acting, record the prediction, owner, review date, and evidence that would change the recommendation. At the checkpoint, mark the result confirmed, revised, reversed, or inconclusive instead of rewriting the original rationale.

Council outcome ledger

The 18 lenses

Members are analytical instruments, not impersonation claims. Each persona has a grounding protocol, a method, known blind spots, and a structured response contract.

Council panel composition

Member Primary lens Useful counterweight
Aristotle Categories and structure Lao Tzu on emergence and excess structure
Socrates Assumption destruction Feynman on reconstruction from first principles
Sun Tzu Terrain and adversarial strategy Aurelius on internal control and moral cost
Ada Lovelace Formal systems and abstraction Machiavelli on incentives and informal power
Marcus Aurelius Resilience and moral clarity Sun Tzu on external competition
Machiavelli Power and incentives Ada on formal consistency
Lao Tzu Non-action and emergence Aristotle on explicit categories
Richard Feynman Explanation and empirical debugging Socrates on the premise itself
Linus Torvalds Shipping and maintainability Meadows on system-level consequences
Miyamoto Musashi Timing and decisive action Torvalds on acting before the ideal moment
Alan Watts Reframing and false problems Torvalds on concrete implementation
Andrej Karpathy Empirical ML behavior Sutskever on frontier risk
Ilya Sutskever Scaling and AI safety Karpathy on observation and iteration
Daniel Kahneman Cognitive bias Feynman on explicit causal reasoning
Donella Meadows Feedback loops and high-impact interventions Torvalds on local fixes
Charlie Munger Inversion and model lattices Aristotle on single-system classification
Nassim Taleb Tail risk and fragility Karpathy on smooth empirical trends
Dieter Rams User clarity and restraint Ada on what can be formalized

The repository stores each member contract under agents/. Custom panels can use --members, named triads, or the classic, exploration-orthogonal, and execution-lean profiles documented in SKILL.md.

Multi-provider routing

The detection script checks which supported providers are available, then the coordinator distributes seats across them. Polarity pairs are separated when possible so a single model family does not play both sides of a disagreement.

Council provider routing

Provider path Detection
Native host subagents Available through the active supported client
OpenAI codex executable
Google gemini executable
Ollama ollama executable
NVIDIA NIM NVIDIA_API_KEY
Cursor cursor-agent executable or configured login

Preview routing without running a council:

/council --dry-route --triad decision Should we accept this acquisition offer?

Use --no-auto-route to keep native-host defaults. Use --models with a copy of configs/provider-model-slots.example.yaml for an explicit seat map. Provider failure is reported before the seat falls back to the native host.

Installation reference

./install.sh                         # Claude Code
./install.sh --codex                 # Claude Code + Codex
./install.sh --codex-only            # Codex only
./install.sh --gemini                # Claude Code + Gemini CLI
./install.sh --gemini-only           # Gemini CLI only
./install.sh --opencode              # Claude Code + OpenCode
./install.sh --opencode-only         # OpenCode only
./install.sh --copy-configs          # Include provider-routing templates
./install.sh --dry-run               # Preview writes

Custom target directories are supported through --claude-dir, --codex-dir, --gemini-dir, and --opencode-dir. Run ./install.sh --help for the current contract.

Verify the checkout

./scripts/council-simulation-checklist.sh
./install.sh --dry-run --codex
./install.sh --dry-run --gemini
./install.sh --dry-run --opencode

The checklist validates persona structure, host-protocol parity, routing configuration, execution checkpoints, verdict fields, and installer behavior. Use the demo session pack to exercise full, quick, and duo modes.

Repository map

SKILL.md                    canonical coordinator protocol
SKILL.codex.md              Codex host mirror
SKILL.gemini.md             Gemini CLI host mirror
SKILL.opencode.md           OpenCode host mirror
agents/                     18 grounded persona contracts
configs/                    provider and model-routing examples
demos/                      sample sessions and verdict template
scripts/                    detection, conversion, and validation tools
assets/                     brand system and README visuals
install.sh                  multi-host installer

Protocol changes start in SKILL.md and must remain behaviorally aligned with each host mirror. See CONTRIBUTING.md for the required checks and review contract. Security reports belong in SECURITY.md, not a public issue.

Project links

License

MIT © 2026 0xNyk

Frequently asked questions

Is council-of-high-intelligence free to use?

council-of-high-intelligence is open source under the MIT licence. There is no licence fee and no seat count — you can self-host it or, where the project offers one, pay a vendor for a managed version instead.

What does council-of-high-intelligence do?

Structured multi-perspective deliberation for hard decisions. Run full councils, focused triads, or duo debates across Claude Code, Codex, Gemini CLI, and OpenC

What is council-of-high-intelligence written in?

council-of-high-intelligence is primarily written in Shell. Its source is publicly available at https://github.com/0xNyk/council-of-high-intelligence, and it has 4,740 GitHub stars.