Skip to main content

​
Installation

The Electric Monks Dialectic Skill works with any coding agent that supports subagent spawning and web search — Claude Code, Cursor, Windsurf, and similar tools.
This is a heavy process by design. Expect 10–15 minutes per round minimum, and plan for at least 3 rounds. This skill needs the best available model — every phase benefits from maximum reasoning capability.

​
Requirements

Your coding agent must support:
  • Subagent spawning — The skill spawns independent Electric Monk sessions in fresh contexts
  • Web search — For research and evidence gathering
  • File operations — Each dialectic generates several output files

​
Directory structure

Create a directory to collect your dialectics. Each dialectic gets its own subdirectory — the skill generates several files for each session: 
dialectics/
├── tanstack-vs-nextjs/
│   ├── context_briefing.md
│   ├── monk_a_output.md
│   ├── monk_b_output.md
│   ├── determinate_negation.md
│   ├── sublation.md
│   └── dialectic_queue.md
├── agent-governance/
│   └── ...
└── career-decision/
    └── ...

​
Generated files explained

  • context_briefing.md — Research synthesis and domain context that both Monks read
  • monk_a_output.md — First Monk’s fully committed position essay
  • monk_b_output.md — Second Monk’s fully committed position essay
  • determinate_negation.md — Structural analysis of the contradiction
  • sublation.md — The synthesis that transforms the question
  • dialectic_queue.md — Tracks explored and unexplored contradictions for future rounds

​
Setup

Create your dialectics directory: 
mkdir -p ~/dialectics
cd ~/dialectics
The dialectic_queue.md file persists across sessions. You can come back to it in a future session and pick up where you left off with unexplored contradictions.

​
Loading the skill

This skill is designed for OpenCode and compatible coding agents that support skill loading.

​
Get the skill

The skill is available on GitHub: 
git clone https://github.com/KyleAMathews/hegelian-dialectic-skill.git

​
For OpenCode

Load the skill using the skills command: 
npx skills add https://github.com/KyleAMathews/hegelian-dialectic-skill
Or if you’ve cloned the repository locally, point to the SKILL.md file: 
/install-skill /path/to/hegelian-dialectic-skill/SKILL.md

​
For other coding agents

If your agent supports custom skills or agent prompts, provide the path to SKILL.md from the cloned repository. Check your agent’s documentation for the specific command.

​
Model selection

Use the strongest available model. Every phase benefits from maximum reasoning capability:
  • Elenctic interview requires subtle probing
  • Monk essays need full conviction and deep reasoning
  • Structural analysis demands cross-domain pattern recognition
  • Synthesis requires genuine creativity
  • Validation requires meta-level reasoning about belief transformation
The skill is particularly effective with models that support extended thinking (Claude Sonnet, GPT-4, or better).

​
Cost expectations

A single dialectic round typically involves:
  • Novel/obscure domain: 150-250K tokens (full research phase)
  • Well-known domain: 50-100K tokens (minimal research)
  • Personal/values domain: 25-75K tokens (interview-heavy, minimal external research)
Plan for at least 3 rounds to reach meaningful insight. The first round is calibration.

​
Next steps

Quickstart

Run your first dialectic

Introduction

Understand how the skill works

Build docs developers (and LLMs) love

Get started for freeTalk to us