RinDig/icm-architect

Claude skill: design any process, idea, or problem into an ICM workspace (folder structure as agent architecture), or restructure an existing folder into one

What it solves

It eliminates the need for complex orchestration code in multi-agent systems by using a structured folder system as the agent's architecture. This allows humans to easily monitor the system's state and allows a single agent to perform the work of a multi-agent framework by reading the right files at the right time.

How it works

The project implements the Interpretable Context Methodology (ICM), which uses numbered folders for sequencing, hierarchy for context scoping, and markdown files for state. It operates in two modes: Build mode, which scaffolds a new workspace based on one of six proven forms (Pipeline, Umbrella, Record library, Knowledge bundle, Context map, or System map), and Restructure mode, which audits and migrates existing folders into an ICM workspace. Every result is validated via a "walk test" to ensure an agent with no memory can orient and act based solely on the files.

Who it’s for

Developers and AI practitioners who want to create interpretable, file-based agent architectures without writing orchestration code, specifically for users of Claude Code or Claude apps.

Highlights

  • Folder-as-Architecture: Replaces code-based orchestration with a structured directory layout.
  • Six Proven Forms: Provides templates for various needs, from production lines (Pipelines) to navigable brains (Knowledge bundles).
  • Walk Test Validation: Ensures the system is structure-dependent rather than memory-dependent.
  • Dual Mode Operation: Can either build a new workspace from scratch or restructure existing repositories and vaults.

Related

  • Project
    michaelshimeles/skills

    A collection of specialized agent skills for Claude Code that automate complex workflows like visual testing, architectural enforcement, and PR refinement.

  • Project
    CronusL-1141/AI-company

    AI Team OS is an open‑source platform that turns a Claude‑style LLM into an autonomous “CEO”. It continuously runs without prompting, orchestrating specialist agents, a task wall, research pipelines, and a two‑layer memory system. The FastAPI + React dashboard shows live workflows, decision traces, and a unified search over all memos. It’s aimed at technical founders or research teams who want an AI‑first, self‑driving company‑like workflow.

  • Project
    RinDig/Interpretable-Context-Methodology

    A methodology and set of patterns for structuring AI agent workflows using a filesystem hierarchy instead of complex orchestration frameworks, enabling human-in-the-loop review and high observability.

  • Project
    Mathews-Tom/armory

    A curated collection of production-grade skills, agents, and workflows for AI coding agents like Claude Code, providing standardized automation for the full software development lifecycle.

  • Project
    fengshao1227/ccg-workflow

    CCG‑Workflow is a Node‑JS CLI that makes Claude Code the central orchestrator for multiple code‑generation models (Codex, Gemini, etc.). It injects JavaScript hooks to keep context, creates persistent task folders, runs built‑in strategies (e.g., full‑collaborate), applies security/quality gates, and can even deploy to Baota panels. Install with `npx ccg-workflow` (full mode) or as a Claude plugin for just the reusable skills.