Content
# Tool List
[](LICENSE)
[](CHANGELOG.md)
[-lightgrey.svg)]()
[](https://github.com/TserenTserenov/FMT-exocortex-template/tree/en-draft)
<img src="https://raw.githubusercontent.com/TserenTserenov/FMT-exocortex-template/main/docs/assets/orz-cycle.svg" alt="Open, Work, Close — the same cycle at every scale: session, day, week" width="100%">
> Intellectual Work Environment. Your knowledge. Your experience. Your environment — works on top of any AI platform.
>
> **Repository type:** `Base/Formats` (FMT) — distribution template. After forking, it becomes your personal environment with AI agents.
## Problem
AI assistants can generate text, code, and answers. But most users face the same issues:
- **Context is lost.** Each new session with AI starts from scratch. Yesterday's solutions, plans, and agreements are forgotten.
- **Knowledge stays in your head.** You've taken a course, read a book, or solved a problem, but can't reproduce the thought process a month later.
- **AI replaces thinking, not enhances it.** You get an answer but don't become more competent. Without AI, you're back to square one.
- **No system.** Plans are in notes, tasks are in your head, and knowledge is in chats. Everything is disjointed.
- **Time leaks away.** It's unclear what you've worked on, what you've accomplished, or where you're headed.
## Solution: IDE, but for Thinking
**IWE (Intellectual Work Environment)** — an intellectual work environment.
Just like an IDE combines an editor, compiler, and debugger into one environment for programmers, IWE combines knowledge, planning, and AI agents into one environment for thinking.
| IDE (for code) | IWE (for thinking) |
|---------------|-------------------|
| Editor → write code | Exocortex → capture knowledge |
| Compiler → checks syntax | Principles → check solution correctness |
| Debugger → finds errors | ORZ protocols (Open→Work→Close) → find knowledge and context losses |
| Linter → improves quality | ArchGate → evaluates architectural decisions |
| Git → change history | Strategist → work history and planning |
> **Key principle: exoskeleton, not prosthesis.** IWE enhances your thinking, not replaces it. After each session, you become more competent, not just get a result. Read more: [principles-vs-skills.md](docs/principles-vs-skills.md).
## Key IWE Terms
| Term | What is it |
|--------|---------|
| **Exocortex** | Your external memory — files with plans, context, conclusions, which Claude reads in each session |
| **Pack** | Formalized knowledge base in your domain — the single source of truth for domain knowledge |
| **ORZ** | Open → Work → Close — a ritual for each session and each day, prevents context loss |
| **ArchGate** | Structured evaluation of architectural decisions based on 7 characteristics (instead of "it seems good") |
| **Strategist** | AI agent that automatically creates day/week plans and tracks progress |
Full glossary: [ONTOLOGY.md](ONTOLOGY.md)
## Work Culture — a New Style of Interaction with AI
IWE is not a set of prompts. It's a **work culture**: 14 elements (protocols, skills, formats) that turn chaotic interaction with AI into a manageable process. Read more: [DP.M.008 — Operating Rules of IWE](https://github.com/TserenTserenov/PACK-digital-platform/blob/main/pack/digital-platform/03-methods/DP.M.008-iwe-operating-rules.md).
### ORZ Protocol (Open → Work → Close)
Each session and each day go through three stages:
- **Open** — Claude reviews the plan, defines the task, and agrees on the approach. You don't start work "from scratch" — AI knows the context.
- **Work** — as you work, Claude captures valuable knowledge (Capture-to-Pack). Insights are not lost.
- **Close** — the result is recorded, the plan is updated, and the next session starts from where you left off.
Skipping Open = unplanned work. Skipping Close = lost results.
### Exocortex — External Memory
Your knowledge, principles, distinctions, plans, and context are stored in files that Claude reads in each session. This is not a "prompt" — it's an **accumulated base** that grows with you.
### Knowledge Formalization (Pack)
What you've learned is not retained in your head. Valuable knowledge is formalized in Pack — a domain knowledge passport. Pack is the single source of truth for domain knowledge. Read more: [LEARNING-PATH.md](docs/LEARNING-PATH.md).
## For Whom
Every professional is overwhelmed with information: 12+ tools (Notion, Google Docs, Slack, ChatGPT, courses...), knowledge is scattered, and nothing is connected. AI answers questions but doesn't know *your* context — every time from scratch.
IWE is for those who want to change this:
- **Entrepreneurs and managers** — you strategize, make decisions, and manage projects. IWE provides a system: from weekly planning to formalizing domain knowledge
- **Engineers and developers** — you work with code and architecture. IWE retains context between sessions, AI knows your codebase, technical debt, and roadmap
- **Researchers and analysts** — you study, synthesize, and publish. IWE turns scattered notes into a structured knowledge base that grows with you
- **Everyone who does intellectual work** — and wants **symbiosis with AI**, not dependence on it. Exoskeleton for thinking, not prosthesis
## Use Cases
### Work Projects
| Scenario | What happens | Read more |
|----------|---------------|-----------|
| **Product development** | Claude knows architecture, technical debt, and roadmap. Each session is a continuation, not a new start | [SC.013](docs/use-cases/USE-CASES.md#sc013-working-session-with-claude-code), [SC.015](docs/use-cases/USE-CASES.md#sc015-system-development-ds) |
| **Documentation management** | Knowledge is captured in Pack as you work. No need to "write docs later" — they are written during work | [SC.004](docs/use-cases/USE-CASES.md#sc004-knowledge-capture-and-extraction), [SC.014](docs/use-cases/USE-CASES.md#sc014-knowledge-formalization-pack) |
| **Project coordination** | WeekPlan, DayPlan, working product registry — Strategist helps plan and track progress | [SC.001](docs/use-cases/USE-CASES.md#sc001-day-planning), [SC.002](docs/use-cases/USE-CASES.md#sc002-weekly-planning-and-review) |
| **Review and refactoring** | ArchGate evaluates decisions based on 7 characteristics. Not "it seems good," but a structured assessment | [SC.015](docs/use-cases/USE-CASES.md#sc015-system-development-ds) |
### Personal Development
| Scenario | What happens | Read more |
|----------|---------------|-----------|
| **Taking a course** | Claude helps capture key ideas, asks questions to check understanding, and connects new knowledge to existing knowledge | [SC.003](docs/use-cases/USE-CASES.md#sc003-learning-and-development) |
| **Writing articles** | Creative pipeline: note → draft → template → publication. Each artifact is tracked | [SC.005](docs/use-cases/USE-CASES.md#sc005-publishing-content) |
| **Strategizing** | Weekly session: review of the previous week, planning for the new week, and checking goals. Strategist prepares a draft — you make decisions | [SC.011](docs/use-cases/USE-CASES.md#sc011-strategizing) |
| **Building a knowledge base** | Your Pack grows. In six months, you have a formalized domain knowledge base, not a collection of notes | [SC.014](docs/use-cases/USE-CASES.md#sc014-knowledge-formalization-pack) |
> Full catalog of 15 scenarios: **[USE-CASES.md](docs/use-cases/USE-CASES.md)**
## How it Looks in Practice
- In the morning, Strategist creates a plan: notification in Telegram + DayPlan file in the repository
- You open VS Code → `claude` → Claude knows what's in the plan and suggests starting with the priority task
- You work — Claude captures knowledge as you go (Capture-to-Pack)
- You close the session — the result is recorded, and the plan is updated
- On Monday, Strategist prepares a draft of the weekly plan, and you discuss it in a strategizing session
## Getting Started
**Quick Start** (Git, Node.js, Claude Code already installed): **[QUICK-START.md](docs/QUICK-START.md)** -- 15 minutes to the first session.
**Full Installation** from a clean computer: **[SETUP-GUIDE.md](docs/SETUP-GUIDE.md)** -- 30-60 minutes with installation of all dependencies.
**Not on macOS or not Claude Code?** Read **[PORTABILITY.md](docs/PORTABILITY.md)** — instructions for Kimi Code, Hermes Agent, and others.
**Another agent or LLM?** IWE is not tied to Claude. If your agent sees files in the repo folder and can edit files, it will work. How to connect → [PORTABILITY.md](docs/PORTABILITY.md).
```bash
mkdir -p ~/IWE && cd ~/IWE
gh repo fork TserenTserenov/FMT-exocortex-template --clone
cd FMT-exocortex-template
bash setup.sh
```
After installation:
```bash
cd ~/IWE
claude
```
Tell Claude: **"Let's have the first strategic session"** -- and it will guide you through goal setting, creating the first plan, and environment setup.
## Customization
IWE updates like a distribution — you get platform updates without losing your settings.
**Extensions (extensions/)** — add your blocks to protocols:
```bash
# Add reflection at the end of the day
echo "## Day Reflection
- What was challenging?
- What would you do differently?
- What to praise?" > extensions/day-close.after.md
```
**Parameters (params.yaml)** — enable/disable protocol steps:
```yaml
reflection_enabled: true # Enable reflection
video_check: false # Disable video check
multiplier_enabled: true # IWE multiplier
```
**Updates** — `bash update.sh` updates the platform, preserving your extensions/, params.yaml, and edits in CLAUDE.md (3-way merge).
Read more: [extensions/README.md](extensions/README.md)
## Documentation
| Document | What's inside |
|----------|-----------|
| **[Beginner's Guide](docs/onboarding/onboarding-guide.md)** | Start here if you're new to IWE. What it is, why, and what's it made of -- without technical terms |
| **[Quick Start](docs/QUICK-START.md)** | 15 minutes from `git clone` to the first session. For those who already have Git and Claude Code |
| **[SETUP-GUIDE.md](docs/SETUP-GUIDE.md)** | Step-by-step installation from a clean computer. Requirements, modes (core/full), verification |
| **[LEARNING-PATH.md](docs/LEARNING-PATH.md)** | IWE learning path: architecture, principles, protocols, Pack, roles |
| **[DATA-POLICY.md](docs/DATA-POLICY.md)** | Data policy: what's collected, where it's stored, how to delete |
| **[IWE-HELP.md](docs/IWE-HELP.md)** | Quick reference and FAQ |
| **[principles-vs-skills.md](docs/principles-vs-skills.md)** | Why principles are more important than skills: generative hierarchy |
| **[ONTOLOGY.md](ONTOLOGY.md)** | Ontology: all terms and abbreviations of IWE |
| **[CHANGELOG.md](CHANGELOG.md)** | Template change history |
## FAQ
**Q: Is an Anthropic subscription required?**
A: For a full installation (Claude Code) — Claude Pro ($20/month) is recommended. If necessary, you can switch to Claude Max (~$100/month) for unlimited work. For minimal (`setup.sh --core`) — works with any AI CLI. Read more: [SETUP-GUIDE.md](docs/SETUP-GUIDE.md).
**Q: Does it work with other AI (not Claude)?**
A: Yes, three agents are supported out of the box:
- **Claude Code** — full support: reads `CLAUDE.md`, all skills and hooks work.
- **Kimi Code** (VS Code) — reads `AGENTS.md` automatically when opening the repo. Customization: `extensions/` or `AGENTS-agent-blocks.md`. Skills (`/day-open`, etc.) through Claude Code.
- **Hermes Agent** — connect Aisystant MCP through Hermes settings, and it will receive instructions automatically.
For other agents (Cursor, Copilot, Gemini), adaptation is required. Read more: [PORTABILITY.md](docs/PORTABILITY.md).
Minimal installation (`setup.sh --core`) works without binding to a specific agent.
**Q: Does it work on Linux/Windows?**
A: Yes. The core works on any OS. Strategist automation: macOS — launchd, Linux — cron, Windows — WSL. Read more: [SETUP-GUIDE.md](docs/SETUP-GUIDE.md).
**Q: What if the computer is turned off or sleeping — will the automation stop?**
A: Cloud Scheduler (GitHub Actions) works in the cloud even when the computer is turned off. For local agents: scripts automatically prevent sleep during work (macOS: `caffeinate`, Linux: `systemd-inhibit`). For laptops, it's recommended to set up automatic wake-up and prohibit idle sleep — see [SETUP-GUIDE.md](docs/SETUP-GUIDE.md).
**Q: What is Pack?**
A: A formalized domain of knowledge — the only source-of-truth for domain knowledge. Read more: [LEARNING-PATH.md](docs/LEARNING-PATH.md).
**Q: Are my data safe?**
A: Three zones of protection: local, GitHub (private repo), platform (per-user isolation). Read more: [DATA-POLICY.md](docs/DATA-POLICY.md).
**Q: How does IWE differ from Obsidian / Notion / Logseq?**
A: Obsidian — a note storage. IWE — **working environment** with protocols, AI agents, and knowledge formalization. You can use Obsidian within IWE for notes, but IWE provides structure, planning, and competency accumulation.
**Q: Do I need to program?**
A: No. The template is a ready-made configuration. Installation via setup.sh. Work — through Claude Code in natural language.
**Q: Can I do without a Strategist?**
A: Yes. Claude Code + CLAUDE.md + memory/ work fully. Strategist — planning automation. Without it, you plan manually.
**Q: How to set up a strategy day?**
A: In `memory/day-rhythm-config.yaml`, change `strategy_day: sunday` to the desired day. Read more: [LEARNING-PATH.md](docs/LEARNING-PATH.md).
**Q: Cloning ended up in `~` instead of `~/IWE`?**
A: All installation commands must be executed in one terminal. If you opened a new one — it starts from `~`. Delete the folder from `~` and repeat from `cd ~/IWE`. Read more: [SETUP-GUIDE.md](docs/SETUP-GUIDE.md).
> Full FAQ with detailed answers: [DP.IWE.002 §11](https://github.com/TserenTserenov/PACK-digital-platform/blob/main/pack/digital-platform/02-domain-entities/DP.IWE.002-iwe-template-and-setup.md#11-faq). Practical guide: [LEARNING-PATH §11](docs/LEARNING-PATH.md#11-quick-reference).
## IWE Community
IWE is an environment that you build alone. But you develop it together.
**IWE Community** — a closed chat for practitioners who work with the same system: ORZ, Pack, exocortex. A place where they discuss not "how to prompt better" but how to build serious intellectual work.
### What's happening there
- **Analysis of working products** — participants show real Packs, plans, retrospectives. Get feedback from people who understand what "closing without fixing the result" means.
- **Experience with installation and customization of IWE** — what broke, how to fix, which extensions were useful.
- **Discussion of methods** — ORZ-fractal, ArchGate, Capture-to-Pack in practice: what works, where theory diverges from reality.
- **Links and finds** — tools, patterns, SOTA that fit into the IWE philosophy.
### Why it's important
You can study the system alone. But most questions arise at the application stage: "How to formalize this domain of knowledge?", "Am I using ORZ correctly?", "Who has experience with this tool?"
In the community, these questions get answers from people who have already gone through it.
### Free channels
- [GitHub Discussions](https://github.com/TserenTserenov/FMT-exocortex-template/discussions) — questions, ideas, show your setup.
- [Issues](https://github.com/TserenTserenov/FMT-exocortex-template/issues) — bug reports and feature requests.
### Closed community (Telegram)
In-depth practice, analysis of working products, direct support. Entry — through the seminar **"IWE for practitioners"** (5000₽) in the bot [@aist_me_bot](https://t.me/aist_me_bot).
## Contribution to development
See [CONTRIBUTING.md](CONTRIBUTING.md) — how to help the project.
**IWE team developers (T4+ level):** a single entry point — [Where to start as a developer](docs/developer/). In 10 minutes, you'll understand the development pipeline (6 stations, double output) and make your first task.
## License
MIT
Connection Info
You Might Also Like
No recommendations available