council-of-high-intelligence — What is it?

The Council of High Intelligence is an AI-driven decision-making tool that facilitates structured, multi-perspective deliberation for complex decisions across various LLM providers.

⭐ 3,449 Stars 🍴 307 Forks Shell MIT Author: 0xNyk
Source: Official README and linked documentation View on GitHub →

Why it matters

This project is being discussed because its innovative approach to decision-making by use multiple AI providers and historical thinkers' perspectives. It addresses the pain point of subjective decision-making by providing diverse viewpoints and structured analysis. The unique technical choice of incorporating 18 historical thinkers as AI personas differs.

Source: Official README and linked documentation

Core Features

Multi-provider routing

The project supports routing across multiple AI providers, ensuring diverse perspectives and reducing reliance on a single model family.

Source: Official README and linked documentation
Structured multi-round deliberation

It employs a structured protocol for deliberation, including problem restatement, analysis, cross-examination, and synthesis, leading to a confident-weighted verdict.

Source: Official README and linked documentation
18 AI personas

The project uses 18 historical thinkers as AI personas, each with a unique grounding protocol and method, providing diverse analytical perspectives.

Source: Official README and linked documentation

Architecture

The architecture is modular, with distinct components for persona contracts, provider routing, and the main deliberation protocol. It employs a structured data flow, with clear separation of concerns between different phases of the decision-making process.

Source: Official repository tree and dependency files

Project Knowledge Graph

Knowledge graph: project (center) + core features (inner hexagons) + key dependencies (outer chips) Not enough information.Not enough inf… Multi-provider routingMulti-provider rout… Structured multi-round deliberationStructured multi-ro… 18 AI personas council-of-high-inte… Project Core feature Key dependency

Tech Stack

LanguageShellFrameworkClaude Code, Codex, Gemini CLI, OpenCode
Not enough information.
Not enough information.
Source: Official README and linked documentation

Quick Start

To install the plugin for Claude Code, run: /plugin marketplace add 0xNyk/council-of-high-intelligence /plugin install council@council-of-high-intelligence For Codex, Gemini CLI, or OpenCode, clone the repository 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 Restart the target client after installation.
Source: Official README and linked documentation

Use Cases

The project is suitable for teams or individuals facing complex decisions that require diverse perspectives and structured analysis. It is useful in scenarios such as strategic planning, product development, and ethical decision-making.

Source: Official README and linked documentation

Strengths & Limitations

Strengths

  • Provides diverse perspectives on complex decisions.
  • Structured protocol ensures thorough analysis.
  • Supports multiple AI providers for analysis.

Limitations

  • Limited information on key dependencies.
  • May require technical expertise to set up and use effectively.
Source: Official README and linked documentation

Latest Release

The latest formal release is v1.2.0, published on 2026-07-04. It added a Claude Code marketplace, confidence-weighted voting, role reasoning_method values, project-level.council.yaml, and Cursor CLI routing, and fixed host-protocol drift.

Source: github.com/0xNyk/council-of-high-intelligence/releases/tag/v1.2.0

Verification and evidence

status: docs-only  |  checked_at: 2026-09-23  |  environment: Docs only. Reviewed the official README, SKILL.md, install.sh, plugin manifest, routing configuration, MIT license, and v1.2.0 release. The council was not installed or run.  |  limitations: Installation writes, provider detection, parallel calls, anonymization, vote calculations, cost, fallbacks, prompt-injection defenses, and verdict quality were not verified.  |  observed_output: The official Release API returns v1.2.0 dated 2026-07-04, and the plugin manifest also reports 1.2.0. The README gives installation paths for Claude Code, Codex, Gemini CLI, and OpenCode.

Source: github.com/0xNyk/council-of-high-intelligence/releases/tag/v1.2.0 (1) · github.com/0xNyk/council-of-high-intelligence/blob/main/.claude-plugi… (2)

Comparison

Karpathy's llm-council is a local web app that collects answers from several models through OpenRouter, runs anonymous peer review, and asks a chair model to synthesize the result. Its README calls it a one-off experiment with no maintenance plan. Council of High Intelligence is a multi-CLI skill centered on role methods, cross-examination, voting, and action templates. llm-council is more direct for viewing model answers side by side; Council of High Intelligence better fits a decision process embedded in coding CLIs.

Source: github.com/0xNyk/council-of-high-intelligence (1) · github.com/karpathy/llm-council (2)

Verdict

The Council of High Intelligence is a tool for enhancing decision-making processes by incorporating diverse AI perspectives. It is particularly suited for teams that value structured analysis and diverse viewpoints in their decision-making.

Frequently Asked Questions

What did the documentation review verify for council-of-high-intelligence on 2026-09-23?

The official Release API returns v1.2.0 dated 2026-07-04, and the plugin manifest also reports 1.2.0. The README gives installation paths for Claude Code, Codex, Gemini CLI, and OpenCode.

What evidence boundary applies to this council-of-high-intelligence review?

The review environment was: Docs only. Reviewed the official README, SKILL.md, install.sh, plugin manifest, routing configuration, MIT license, and v1.2.0 release. The council was not installed or run. council-of-high-intelligence behavior outside that scope still needs a local check.

Sources and review scope
This review is based on the project's public documentation. Versions, installation commands, and feature boundaries can change upstream; check the cited sources and current release before use. Sources checked on: 2026-09-23 10:52.

Reference materials: README, GitHub API, dependency files