🎉 NullShift is live
The first public release of NullShift — an AI-powered triage assistant for L1 SOC analysts.
What's in the box
- 🧠 12 LLM providers — Claude Agent SDK (use your Claude subscription, no API key), cloud APIs (Anthropic, OpenAI, Gemini, Groq, xAI, DeepSeek, Perplexity, OpenRouter, Qwen, Kimi), or fully local Ollama — even hosted on another machine over Tailscale.
- 🔌 5 SIEM connectors — Wazuh, LimaCharlie, Splunk, Elastic, Microsoft Sentinel.
- 📚 RAG over 753+ playbooks — Anthropic-Cybersecurity-Skills corpus bundled (skills + MITRE ATT&CK mappings), plus 4 NullShift-specific top-level playbooks.
- 📋 Structured investigation reports — SECTION 1 (evidence) → SECTION 2 (reasoning) → SECTION 3 (verdict + confidence).
- 🎯 L1 → L2 handoff mode — ticket-ready escalation summaries on demand.
- 🔧 Admin UI with live SIEM test, drag-and-drop LLM provider chain, per-user temperature slider, conversation search, verdict badges, debug traces.
- 💻
nullshiftCLI —start / stop / restart / status / logs / setup. - ⚡ Single-command setup —
python setup.pyhandles venv, dependencies, JWT, admin account, SIEM, RAG, and starts the server.
Connector maturity
| Status | Connector |
|---|---|
| ✅ Production-ready | Wazuh, Splunk |
| 🧪 Beta — under active testing | LimaCharlie, Elasticsearch, Microsoft Sentinel |
Get started
git clone https://github.com/hegazi-sec/nullshift.git
cd nullshift
python setup.pyThen open http://localhost:58443.
Credits
Knowledge base powered by Anthropic-Cybersecurity-Skills by Mahipal (mukul975) — Apache 2.0.
— Ahmed Hegazi