Features

What Makes NyxStrike Stand Out

NyxStrike is the only open-source platform that combines an AI-driven attack-chain planner, 185+ real security tools, and a full operator workbench — all accessible from a single MCP connection or web dashboard.

Feature	Description
Built-in AI Assistant	Chat widget available on every page — start a conversation without leaving your workflow. Powered by your configured LLM (Ollama, OpenAI, or Anthropic)
Sessions & Operator Workbench	Every engagement lives in a structured session with 4-tab workbench (Workflow, Findings, Notes, Timeline), artifact chaining, AI analysis, and report generation. 7 session creation modes including intelligence-planned, manual, and 4 AI-driven variants
185+ Tools, One Interface	Network recon, web exploitation, WiFi pentesting, binary analysis, cloud auditing, OSINT, password cracking, forensics, and more — all orchestrated through a single MCP connection
Intelligent Attack-Chain Engine	Catalog-driven planner with contextual learning, smarter tool ranking, attack-chain preview with tool selection reasons, and user-selectable precision (quick, comprehensive, stealth) before a session starts
Artifact Chaining	Step-by-step tool chaining with operator approval, confidence hints, and output mapping between steps — not just fire-and-forget
Durable Session Handoff	AI planning and manual dashboard execution share the same session state, so you can switch between agent-driven and hands-on control at any point
Purpose-Built AI Agents	Standalone agents for bug bounty workflows, CTF solving, CVE intelligence, exploit generation, vulnerability correlation, technology detection, rate-limit detection, and intelligent failure recovery
9 Workflow Skills	Pre-built playbooks (web recon, SMB enum, binary analysis, subdomain enum, cloud audit, exploitation, password cracking, and more) mounted as MCP resources the agent follows automatically
Built-in Dashboard	KPI cards, resource monitoring, tool availability by category, live process management (pause/resume/terminate), session workbench, notes, findings tracker, and report generation. No terminal required after setup
Global Command Palette	Ctrl/Cmd+K keyboard-first control: jump pages, trigger tools, compare previous runs
Persistent Run History	Server-side history survives browser refresh. Clears safely with confirmation
Per-Tool Timeout Policies	Inactivity watchdog and max runtime cap keep long-running scans from hanging
Theme System	One-click switcher with live preview: Dark Ops, Candy Pop, Solarized, Unicorn, Forest, and more
LLM Flexibility	Works with Ollama (local), OpenAI, and Anthropic out of the box
Performance Modes	--compact for lightweight/local LLM usage, --profile for targeted tool loading