Skip to content

Home

Jump to bottom
kanadhiayash edited this page Jun 10, 2026 · 4 revisions

Zeref OS Wiki

Zeref OS

Imagine you are an architect working on a major building. Every morning a different contractor shows up. Before they can lay a single brick, you have to re-explain the blueprint, the constraints, the decisions you and the prior contractor made, and what's already been built. Every conversation starts from zero.

That is what working with AI assistants is like today. Each new session — Claude, Codex, Gemini, Cursor, Aider — starts blind. Context evaporates the moment the window closes.

Zeref OS is the persistent memory layer that fixes this. A per-project markdown wiki that AI sessions read first, write to safely, and hand off cleanly. Your project memory travels with the project — not the tool.

Current version: v2.6 · CHANGELOG · Notion Command Center

Placeholder: assets/poc-stateless-vs-zeref.png — add screenshot of two sessions, before / after Zeref OS.

Quick links

  • 📦 Installation — per-harness setup, verification, uninstall
  • 🏗️ Architecture — 6 agents · 14 skills · 8 commands · 6 team packs · 4 Auto-Activation Gates · Model-Tier Routing
  • 🧠 Memory-Model — flat layout, boundary-first reads, contradiction handling, PATTERNS.jsonl event schema
  • 🔒 Privacy-Model — PRIVACY / REDACT / SHARING_POLICY, modes, connectors, R6 Zero Context Loss
  • 👥 Team-Packs — solo / build / research / red / audit / ship
  • 🔍 Pattern-Detection — Two-Strikes Rule, pattern-observer, skill drafting
  • ❓ FAQ — common questions
  • 📖 Glossary — boundary file, evidence grade, Two-Strikes Rule, R6, 4-gate chain, model-resolver
  • 🌱 Inspirations — engineering lineage and influences

What v2.6 ships

Per-project flat memory/ wiki in plain markdown · append-only PATTERNS.jsonl event log with schema validator · point-in-time snapshots · contradiction safety with human arbitration · three privacy modes (exact / abstract / local-only) · six on-demand team packs · cross-harness handoff format with caveman-grammar compression · and a 4-gate auto-activation chain that classifies every major task on cost weight, picks the smallest useful skill stack, probes extended-tool reachability, and restructures unstructured prompts before any token spend.

The 4-gate chain (every major task)

[budget-governor]       classify weight (CRITICAL/HIGH/MEDIUM/LOW) + match model tier
       ↓
[skill-router]          pick smallest stack (1 lead + 2-3 support + 1 QA, max 5 skills)
       ↓
[fleet-activator]       live-probe ECC / claude-obsidian / Graphify / browser-harness / notebooklm / gstack
       ↓
[prompt-context-engine] classify STRUCTURED / SEMI-STRUCTURED / UNSTRUCTURED; restructure if needed; R6 zero context loss
       ↓
                        execute (declared stack, declared brief, declared tier)
       ↓
[caveman-handoff]       compress cross-model handoff (40-60% reduction; NFKC + R6 diff)

Each gate declares its result inline. User can override before token spend. Per AGENTS.md ## Auto-Activation Gates.

Where to start

If you want to… Read
Install in 5 minutes Installation
Understand the system Architecture → Memory-Model
Lock down privacy first Privacy-Model
See how teams work Team-Packs
Understand who we built on Inspirations
Run the validator yourself scripts/zeref-validate.py

README · AGENTS.md · CHANGELOG · GITHUB_OS.md · _shared/model-resolver.md

Zeref OS

Getting Started

Architecture

Reference

Clone this wiki locally