JEBAT v8.2.2 — Sovereign Agent OS & Agent Workstation

Sovereign execution, private memory, and audited intelligence.

JEBAT is an enterprise-grade self-hosted AI platform and agent workstation. It provides governed local LLM inference, secure cognitive routing, multi-agent swarm orchestration, embedded threat reconnaissance, and eternal memory — all running fully air-gapped on your private network with zero data leakage.

Named after the legendary Malay warrior Hang Jebat — loyal, powerful, and unforgettable.

🚀 Quick Start

Zero-Config Installation (npx / bunx)

# Run immediately — bootstraps the JEBAT CLI on first run
npx @nusabyte/jebat repl

# Or with bun
bunx @nusabyte/jebat repl

# One-shot chat
npx @nusabyte/jebat chat "Explain the memory consolidation algorithm"

# Run agent task
npx @nusabyte/jebat agent "Audit all API endpoints in src/services"

Requirements

| Tool | Version | Install | |------|---------|---------| | Node.js | ≥ 18 | nodejs.org | | Python | ≥ 3.11 | python.org | | pip | Latest | Included with Python |

📦 Installation Methods

Method 1: npx (Recommended — Zero Install)

npx @nusabyte/jebat repl

• Bootstraps the supported JEBAT CLI on first run
• Requires Node.js 18+, Python 3.11+, and pip
• Does not use the unrelated jebat package on PyPI

Method 2: Global Install

npm install -g @nusabyte/jebat
# or
bun install -g @nusabyte/jebat

jebat repl

Method 3: Source Build (Advanced)

git clone https://github.com/nusabyte-my/jebat-core.git
cd jebat-core
curl -fsSL https://raw.githubusercontent.com/nusabyte-my/jebat-core/main/install.sh | bash   # bootstrap the Python core
jebat repl

The npm package is a thin launcher. On first run it downloads and runs the bootstrap, which provisions Python, the venv, and the jebat launcher in ~/.local. Do not run pip install jebat — that name is an unrelated PyPI package (a website checker) and is not JEBAT. The bootstrap is the only supported install path.

🎯 Core Commands

| Command | Description | |---------|-------------| | jebat repl | Interactive REPL — streaming, tools, history (primary interface) | | jebat chat "prompt" | One-shot chat with tool calling | | jebat agent "task" | Run one-shot agent with tool-calling | | jebat config show\|set\|reset\|edit | Full configuration management | | jebat file read\|write\|patch\|search\|undo\|tree | Safe file ops with backups | | jebat tools list\|inspect | Inspect registered tools | | jebat memory store\|search\|stats | 5-layer eternal memory | | jebat status | System health & provider status | | jebat --version | Show version | | jebat --help | Show help |

🔧 Configuration

Config file: ~/.jebat/jebat.yaml

agent:
  default_model: "qwen2.5-coder:7b"
  default_provider: "ollama"
  safety_mode: "confirm"
  max_iterations: 10
  stream_tokens: true

llm_providers:
  ollama_host: "http://localhost:11434"
  fallback_providers: ["openrouter", "groq", "openai"]

security:
  enable_guardrails: true
  audit_log: true

Quick Config Commands

# Show config
npx @nusabyte/jebat config show

# Set value
npx @nusabyte/jebat config set agent.safety_mode confirm

# Edit in $EDITOR
npx @nusabyte/jebat config edit

🏗️ Architecture

┌─────────────────────────────────────────────────────────────┐
│  JEBAT Wrapper (this package)                                │
│  • Auto-installs Python package via pip                      │
│  • Forwards CLI arguments                                    │
│  • Handles stdin/stdout for REPL                             │
└───────────────────────┬─────────────────────────────────────┘
                        ▼
┌─────────────────────────────────────────────────────────────┐
│  JEBAT Python Core (via bootstrap — NOT pip install jebat)    │
│  • ReAct AgentLoop (Think→Act→Observe)                       │
│  • Ultra-Think (7 reasoning modes)                           │
│  • Ultra-Loop (autonomous agent)                             │
│  • 5-Layer Memory (M0→M4 with heat scoring)                  │
│  • Registered tools and workspace integrations               │
│  • MCP Client + Server                                       │
│  • Multi-Agent Swarms (Tukang/Hulubalang/Pawang)             │
└─────────────────────────────────────────────────────────────┘

🔐 Security & Privacy

• 100% Local — No data leaves your machine
• Encrypted Storage — Fernet (API keys) + Argon2id (passwords)
• Audit Logging — All tool executions tracked
• 3-Tier Safety — auto / confirm / dangerous
• Sandbox Isolation — Docker for untrusted code
• No Telemetry — Opt-in only

🤝 MCP Integration

The npm launcher starts the JEBAT CLI. MCP server hosting requires a full JEBAT workspace checkout and the jebat-mcp entry point; see the MCP guide. Do not rely on an unauthenticated shared public MCP endpoint.

# From a full workspace checkout
python ./jebat-mcp --transport stdio

# Or self-host HTTP behind an authenticated reverse proxy
python ./jebat-mcp --transport http --host 127.0.0.1 --port 8099

🌍 Platform Support

| OS | Shell | Status | |----|-------|--------| | macOS | zsh/fish/bash | ✅ Full | | Linux | bash/zsh/fish | ✅ Full | | Windows | PowerShell/CMD/Git Bash | ✅ Full | | WSL2 | bash | ✅ Full |

📚 Documentation

• Full README
• Architecture
• Quick Reference
• Deployment
• Roadmap

🤝 Contributing

# Clone repo
git clone https://github.com/nusabyte-my/jebat-core.git
cd jebat-core/jebat-npm

# Test locally
npm link
jebat repl

📄 License

MIT License — See LICENSE file.

Developed under strict data residency governance by NusaByte.
Built with ❤️ by Shaidan Shaari (humm1ngb1rd).

🙏 Acknowledgments

• Python core: jebat-core
• Ollama — Local LLM inference
• LangGraph — Cognitive orchestration
• Rich / prompt-toolkit — Terminal interfaces
• Playwright — Browser automation

🗡️ The JEBAT Way

"Hang Jebat fought with loyalty and honor. JEBAT remembers with precision and purpose."

Your AI. Your Data. Your Legacy.

🗡️ JEBAT v8.2.2 — Because warriors remember everything that matters.