Norwegian builder · Security researcher · Reverse engineer · Open-source enthusiast
I'm NetVar — founder of hashward and co-founder of BitterSecurity. I work on autonomous security agents, reverse-engineering tooling, and the infrastructure that lets AI agents operate reliably on real systems.
My interests connect ethical hacking, reverse engineering, AI/ML, and smart contract security. Open source is where I share the work and learn from others.
Clear assumptions. Reproducible experiments. Conclusions that stay within the evidence.
| Project | Stars | What it is |
|---|---|---|
| Decepticon | 6.4k | Autonomous hacking agent for authorized red-team operations. The largest public project I contribute to. |
| Vigilo | 70 | AI hacker for Web3 smart contracts — built for bug bounties and audit workflows. |
| PurpleAILAB | 7 | Agentic cybersecurity harness with its marketing site and knowledge vault. |
| AutoBot | 1 | Autonomous defense agent for the blue team. |
| Cybertron | 1 | Autonomous security intelligence for humanity. |
| Project | Stars | What it is |
|---|---|---|
| vibe-island | 70 | Dynamic Island for AI coding tools — Rust + Tauri, macOS / Windows / Linux. |
| claude-for-firefox | 19 | Claude for Firefox — a community browser extension with sidebars. |
| omniwire | 17 | Infrastructure for AI agent swarms — 88 MCP tools, A2A protocol, mesh networking. |
| KEVLAR | 6 | Windows x64 kernel-driver emulation and behaviour analysis, powered by Unicorn + Zydis. |
| decepticon-ghidra-mcp | 5 | Ghidra MCP — P-code, BSim, version tracking, and emulation. |
| mcp-ctl | 3 | Zero-intervention MCP lifecycle manager for every AI CLI. Written in Zig. |
| agent-skill-canon | 2 | 1,200+ executable skill files for coding agents and security harnesses. |
| codex-patcher-cc | 2 | Codex CLI patcher — Rust Mach-O patches, wrapper, and config injection. |
| Project | What it is |
|---|---|
| account-security | Account takeover, credential stuffing, session abuse, and identity fraud for game platforms — graph-based analysis. |
| cheat-intel | Cheat-community intelligence: monitoring methodology, trend reporting, and the intel → detection feedback loop. |
| apex-anticheat-lab | FPS anti-cheat lab — runnable synthetic fixtures, cohort-aware baselines, explainable scores. |
| audit-findings-skills | Audit workflow findings packaged as reusable agent skills. |
recurse — AI-native IDE for reverse engineering · ultrafuzz — agentic smart-contract fuzzing · flipperone-mcu-firmware — Flipper One RP2350 co-processor firmware · linux-rk3576-rocket — RK3576 NPU RFC patch series for mainline · AiDA — native C++ assistant plugin for IDA Pro 9.x
Research scope & limitations
These are independent research projects, not evidence of production deployment or employment with a game studio or anti-cheat vendor.
- The anti-cheat lab uses synthetic data.
- Its YARA sketches are not operational detection rules.
- Decepticon is separate work intended for authorized red-team use.
|
Languages Python · TypeScript · Rust · C++ · C · Go · Zig · Java · JavaScript · Solidity · Assembly · Bash Reverse engineering IDA Pro 9.x + IDAPython · Ghidra · radare2 · r2pipe · Unicorn · Zydis · Capstone · YARA · LIEF |
AI & agents MCP (Model Context Protocol) · agent swarms · A2A · Playwright · Claude Code · Codex CLI · prompt engineering Systems Linux kernel · eBPF · Tauri 2 · React · Node.js · Docker · GitHub Actions · PostgreSQL · Redis Hardware Flipper One · RP2350 · embedded firmware · USB protocol work |
| Focus | Questions that interest me | |
|---|---|---|
| 🛡️ | Security & ethical hacking | Where are the attack paths, and how can authorized testing improve defenses? |
| 🔬 | Reverse engineering | What does the software actually do beyond the source code? |
| 🧠 | AI / ML | Where can models help security research and tooling? |
| 🤖 | Agent infrastructure | How do you make autonomous agents reliable, inspectable, and safe to point at real systems? |
| 🔗 | Smart contracts & Web3 | Which trust assumptions break, and what happens when they do? |
| 🌱 | Open source | How can shared, inspectable work help others build and learn? |
| 01 / Context | 02 / Evidence | 03 / Permission |
|---|---|---|
| Understand the threat model and what the data can establish. | Explain the detection, expose limitations, and make experiments reproducible. | Test within an explicitly authorized scope and disclose responsibly. |
Open-source collaboration · Security research · Detection engineering · Agent infrastructure
Find me on GitHub ↗ · Connect on X ↗ · TryHackMe ↗
Build openly. Think adversarially. Verify the evidence.