WinSentinel.Cli 1.16.1

<div align="center">

🛡️ WinSentinel

Your Always-On Windows Security Agent

Not just an auditor - a living agent that monitors, detects, correlates, and responds 24/7.

Real-time threat detection • 13 audit modules • Auto-remediation • Chat control plane • AI-powered correlation • Compliance profiles

🚀 Quick Start · 📦 Install · 📖 Docs · 🐛 Issues · 📋 Changelog

---

</div>

Why WinSentinel?

Most Windows security tools run once and give you a report. WinSentinel is different:

• Always on - runs as a Windows Service, monitoring your system 24/7
• Correlates events - doesn't just flag individual events, it detects multi-stage attack patterns
• Takes action - auto-remediates threats with full undo support
• Speaks your language - chat-based control plane with 25+ commands plus natural language
• Compliance-aware - built-in profiles for Home, Enterprise, HIPAA, PCI-DSS, and CIS L1
• Deeply tested - 4,173 automated tests across 128 test files

---

🏗️ Architecture

Two-process design: a background agent (Windows Service) and a WPF dashboard connected via named pipe IPC.

graph TB
    subgraph Agent["🛡️ WinSentinel Agent (Windows Service)"]
        PM[⚙️ Process Monitor]
        FM[📁 File System Watcher]
        EL[📋 Event Log Listener]
        NM[🌐 Network Monitor]
        SA[📊 Scheduled Auditor<br/>13 modules]
        AB[🧠 Agent Brain +<br/>Threat Correlator]
        AR[🔧 Auto-Remediator<br/>7 actions + undo]
        IS[🔒 Input Sanitizer]
        IPC_S[📡 IPC Server]
    end

    subgraph Dashboard["💻 WinSentinel Dashboard (WPF)"]
        LD[📊 Live Dashboard +<br/>Score Gauge]
        TF[⚠️ Real-time Threat Feed]
        CC[💬 Chat Control Plane]
        SH[📈 Score History & Trends]
        CP[📋 Compliance Profiles]
        FR[🔕 Finding Rules]
        PC[⚙️ Policy Configuration]
        ER[📤 Export Reports]
    end

    PM & FM & EL & NM -->|events| AB
    SA -->|findings| AB
    AB -->|threats| AR
    AB -->|alerts| IPC_S
    IPC_S <-->|Named Pipe IPC| Dashboard

Threat Detection Flow

flowchart LR
    A[Raw Events] --> B{Monitor<br/>Modules}
    B --> C[Process<br/>Events]
    B --> D[File<br/>Events]
    B --> E[EventLog<br/>Events]
    B --> F[Network<br/>Events]
    C & D & E & F --> G[🧠 Threat<br/>Correlator]
    G -->|Single event| H[Classify &<br/>Score]
    G -->|Multi-event<br/>pattern| I[Attack Chain<br/>Detection]
    H & I --> J{Severity?}
    J -->|Critical/High| K[⚡ Auto-Remediate]
    J -->|Medium| L[⚠️ Alert User]
    J -->|Low/Info| M[📝 Log]
    K --> N[Undo Log]

The agent runs continuously - even with the dashboard closed - watching processes, file changes, event logs, and network activity. When it detects suspicious behavior, it correlates events, classifies threats, and auto-remediates based on configurable policies.

---

⚡ Features at a Glance

Category	What You Get
🔍 4 Real-Time Monitors	Process creation/termination, file system changes, Windows Event Log, network connections - always watching
🧠 AI-Powered Brain	Correlates individual events into attack chains. Detects multi-stage attacks that single-event analysis misses
🔧 7 Auto-Remediation Actions	Kill process, quarantine file, block IP, disable account, restore hosts, re-enable Defender, revert registry - all with undo
💬 Chat Control Plane	25+ commands plus natural language. Run audits, query threats, configure policies - from the chat panel
📊 13 Audit Modules	Firewall, Updates, Defender, Accounts, Network, Processes, Startup, System, Privacy, Browser, App Security, Encryption, Event Log
📋 Compliance Profiles	Home, Enterprise, HIPAA, PCI-DSS, CIS L1 - context-aware scoring with per-profile severity adjustments
🔕 Finding Suppression	Ignore/suppress known-acceptable findings with regex rules, expiration dates, and audit trail
📈 Score History	SQLite-backed audit tracking with trends. See your security posture change over time
📤 Export Reports	HTML, JSON, Text, Markdown - save and share results
🔔 Toast Notifications	Critical finding alerts via Windows notification center
💻 CLI Mode	Full CLI (winsentinel.exe) for scripting, automation, and CI/CD gate checks
⚙️ System Tray	Minimize to tray and run silently in the background
🛡️ Input Sanitization	Centralized security layer prevents command injection in all user-facing inputs

---

📸 Sample Audit Output

╔══════════════════════════════════════════════════════╗
║           WinSentinel Security Audit Report          ║
║              2026-02-21 22:35:00 PST                 ║
║           Profile: Enterprise                        ║
╠══════════════════════════════════════════════════════╣
║                                                      ║
║         Security Score:  92 / 100   Grade: A         ║
║         ████████████████████████████████░░  92%       ║
║                                                      ║
╠══════════════════════════════════════════════════════╣
║  Module           Score   Status                     ║
╠══════════════════════════════════════════════════════╣
║  🔥 Firewall       100    ██████████  PASS           ║
║  🔄 Updates          95    █████████░  PASS           ║
║  🛡️ Defender        100    ██████████  PASS           ║
║  👤 Accounts        100    ██████████  PASS           ║
║  🌐 Network          90    █████████░  PASS           ║
║  ⚙️ Processes        90    █████████░  PASS           ║
║  🚀 Startup          95    █████████░  PASS           ║
║  💻 System          100    ██████████  PASS           ║
║  🔒 Privacy          95    █████████░  PASS           ║
║  🌍 Browser          85    ████████░░  PASS           ║
║  📦 App Security     90    █████████░  PASS           ║
║  🔐 Encryption       80    ████████░░  WARN           ║
║  📋 Event Log        85    ████████░░  PASS           ║
╠══════════════════════════════════════════════════════╣
║  Findings: 65 total | 0 critical | 5 warnings       ║
║  Suppressed: 2 (accepted risk)                       ║
╚══════════════════════════════════════════════════════╝

---

🚀 Quick Start

Prerequisites

• Windows 10 or 11 (x64)
• .NET 8 SDK (for building from source)

Clone, Build & Run

git clone https://github.com/sauravbhattacharya001/WinSentinel.git
cd WinSentinel

# Build
dotnet build WinSentinel.sln -p:Platform=x64

# Run the dashboard
dotnet run --project src/WinSentinel.App -p:Platform=x64

# Run tests (4,173 tests)
dotnet test -p:Platform=x64

Quick Audit (no build needed)

.\RunAudit.ps1

---

📦 Installation

Option 1: MSIX Installer

# Downloads cert, installs MSIX - one command
.\Install-WinSentinel.ps1

Option 2: Windows Service

dotnet build src/WinSentinel.Agent -c Release

# Install (requires Administrator)
.\Install-Agent.ps1 -Install

# Check status
.\Install-Agent.ps1 -Status

Option 3: Build MSIX from Source

cd src\WinSentinel.Installer
.\Build-Msix.ps1
# → dist\WinSentinel.msix

---

🔍 Real-Time Monitors

Monitor	What It Watches	Key Detections
⚙️ Process	Process creation & termination	Suspicious executables, unsigned binaries, temp/download path launches, known-bad names
📁 File System	File create/modify/delete/rename	System directory changes, hosts file tampering, startup folder modifications, suspicious DLLs
📋 Event Log	Windows Security & System logs	Failed logons, privilege escalation, audit policy changes, service installations, account modifications
🌐 Network	Active connections & listening ports	New listeners, known-bad IPs, unusual outbound ports, DNS anomalies

---

📊 The 13 Audit Modules

#	Module	What It Scans
1	🔥 Firewall	All profile states, rule analysis, dangerous port exposure (RDP 3389, SMB 445, Telnet 23)
2	🔄 Updates	Windows Update service, pending updates, last install date, update source config
3	🛡️ Defender	Real-time protection, cloud protection, behavior monitoring, definition age, tamper protection
4	👤 Accounts	Local users, admin audit, password policies, guest account, empty passwords
5	🌐 Network	Open ports, SMB/RDP exposure, LLMNR/NetBIOS poisoning, Wi-Fi security, ARP, IPv6
6	⚙️ Processes	Unsigned executables, suspicious paths, high-privilege monitoring
7	🚀 Startup	Startup programs, scheduled tasks, Run/RunOnce keys, service types
8	💻 System	OS build, Secure Boot, BitLocker, UAC level, RDP config, DEP/NX
9	🔒 Privacy	Telemetry, advertising ID, location tracking, clipboard sync, activity history
10	🌍 Browser	Chrome/Edge settings, dangerous extensions, saved passwords, update status
11	📦 App Security	Outdated software, EOL flagging, installed program analysis
12	🔐 Encryption	BitLocker, EFS, certificate store, TPM status
13	📋 Event Log	Failed logins, suspicious events, audit policy gaps

---

📋 Compliance Profiles

Built-in profiles adjust severity weights and scoring for different security contexts:

Profile	Target Environment	Key Adjustments
🏠 Home	Personal/home use	Relaxed - info-level items don't penalize
🏢 Enterprise	Corporate workstations	Moderate - emphasizes patching, network, accounts
🏥 HIPAA	Healthcare environments	Strict - encryption, audit logging, access control weighted heavily
💳 PCI-DSS	Payment card processing	Strict - network segmentation, firewall, patching critical
🔒 CIS L1	CIS Benchmarks Level 1	Comprehensive - baseline security for all organizations

Switch profiles via the dashboard or CLI to see how your system scores under different compliance frameworks.

---

🔧 Auto-Remediation

7 autonomous response actions, each with full undo:

Action	What It Does	Reversible
Kill Process	Terminates suspicious process	-
Quarantine File	Moves to quarantine directory	✅
Block IP	Creates firewall block rule	✅
Disable Account	Disables compromised account	✅
Restore Hosts	Reverts hosts file to clean state	✅
Re-enable Defender	Turns real-time protection back on	-
Revert Registry	Undoes malicious registry changes	✅

---

💬 Chat Control Plane

25+ commands plus natural language understanding:

> status                    # Agent uptime, active monitors
> threats                   # Recent threat events
> audit                     # Run full 13-module audit
> audit firewall            # Run specific module
> score                     # Current score and grade
> history                   # Score trend over time
> monitor status            # All 4 monitor states
> start monitor process     # Start specific monitor
> policy                    # Show current policies
> set risk tolerance high   # Adjust sensitivity
> quarantine                # List quarantined files
> undo <id>                 # Reverse a remediation action
> journal                   # Agent activity log
> export html               # Export report
> fix all                   # Auto-fix all fixable findings

Natural language works too:

> Why is my network score low?
> What's the most dangerous thing on my system?
> Show me failed login attempts from today

---

💻 CLI Reference

# Full audit
winsentinel --audit

# JSON output for scripting
winsentinel --audit --json

# Specific modules only
winsentinel --audit --modules firewall,network,privacy

# CI/CD gate: fail if score < 90
winsentinel --audit --threshold 90

# Auto-fix everything
winsentinel --fix-all

# Compare last two runs
winsentinel --history --compare

# Show what changed
winsentinel --history --diff

Flag	Description
--audit	Run full security audit
--score	Print score and grade only
--fix-all	Auto-fix all fixable findings
--history	View past audit runs
--json / --html / --md	Output format
--output <file>	Save to file
--modules <list>	Comma-separated module list
--threshold <n>	Fail if score below n
--compare / --diff	Compare runs or show deltas
--summary	Executive security summary (plain-English brief)
--quiet	Score + exit code only

Exit codes: 0 = pass, 1 = warnings, 2 = critical, 3 = error

---

📊 Scoring

Starts at 100, deductions by severity:

Severity	Deduction	Example
🔴 Critical	-15 pts	Real-time protection disabled, firewall off
🟡 Warning	-5 pts	LLMNR enabled, outdated definitions
🔵 Info	-1 pt	Telemetry at default level
✅ Pass	0 pts	Secure Boot enabled, UAC on

Grades: A+ (95+) · A (90-94) · B (80-89) · C (70-79) · D (60-69) · F (<60)

Compliance profiles adjust these weights contextually - a finding that's info-level for Home use might be a warning under HIPAA.

---

🏗️ Project Structure

WinSentinel.sln
├── src/
│   ├── WinSentinel.Core/          # Security audit engine (13 modules)
│   │   ├── Audits/                # Firewall, Network, Defender, etc.
│   │   ├── Models/                # AuditResult, Finding, SecurityReport
│   │   ├── Services/              # AuditEngine, Orchestrator, Scorer
│   │   └── Helpers/               # Shell, PowerShell, Registry, WMI
│   │
│   ├── WinSentinel.Agent/         # Always-on agent (Windows Service)
│   │   ├── Modules/               # 4 real-time monitors
│   │   ├── Services/              # Brain, Correlator, Remediator, Chat
│   │   │                          # Journal, Policy, IPC, Sanitizer
│   │   └── Ipc/                   # Named pipe protocol
│   │
│   ├── WinSentinel.App/           # WPF dashboard (MVVM)
│   │   ├── Views/                 # Dashboard, Chat, Policy, Compliance
│   │   ├── ViewModels/            # CommunityToolkit.Mvvm
│   │   └── Services/              # IPC client, ChatAI
│   │
│   ├── WinSentinel.Cli/           # Command-line interface
│   └── WinSentinel.Installer/     # MSIX packaging
│
├── tests/
│   └── WinSentinel.Tests/         # 4,173 xUnit tests (128 files)
│
├── RunAudit.ps1                   # Quick audit script
├── Install-Agent.ps1              # Service installer
├── Install-WinSentinel.ps1        # MSIX installer
└── Fix-Network.ps1                # Network security fix script

By the numbers: 72k+ lines of source code, 50k+ lines of tests, 370+ commits, 128 test files.

---

⚙️ Tech Stack

Component	Technology
Runtime	.NET 8 (LTS)
UI	WPF + MVVM (CommunityToolkit.Mvvm)
Language	C# 12
Agent	Microsoft.Extensions.Hosting + Windows Services
IPC	Named Pipes (System.IO.Pipes)
Database	SQLite (Microsoft.Data.Sqlite)
Testing	xUnit — 4,173 tests
Security	CodeQL scanning, input sanitization
Packaging	MSIX with code signing
CI/CD	GitHub Actions (build, test, release, CodeQL)
AI	Ollama (local LLM) + built-in rule engine

---

📋 Releases

Version	Date	Highlights
v1.4.2	2026-04-16	Test coverage improvements
v1.4.1	2026-04-02	NuGet packaging & CI fixes
v1.4.0	2026-03-29	Major feature release
v1.3.0	2026-03-20	CLI power tools & hardened security
v1.2.0	2026-03-16	Deep system auditing & threat intelligence
v1.1.0	2026-02-20	Compliance profiles, finding suppression, remediation checklists
v1.0.0	2026-02-17	Initial release — always-on agent, 4 monitors, AI correlator, 13 audit modules

---

🔐 Security Model

WinSentinel follows a defense-in-depth approach:

Layer	Protection
Input Sanitization	All user inputs (chat, CLI, config) pass through a centralized InputSanitizer that blocks command injection, path traversal, and control characters
Least Privilege	The dashboard runs as the current user. Only the agent service and remediation actions require Administrator
Undo Journal	Every auto-remediation action is logged with full undo metadata - quarantined files can be restored, blocked IPs unblocked, disabled accounts re-enabled
Named Pipe IPC	Dashboard↔Agent communication uses local-only named pipes (no network exposure). The pipe is ACL-restricted to the installing user and SYSTEM
Finding Suppression Audit Trail	When you suppress a finding, WinSentinel records who, when, why, and expiration - suppressions don't silently hide real threats
No Outbound Telemetry	WinSentinel sends zero data home. All analysis is local. AI features use local Ollama models only

Reporting vulnerabilities: See SECURITY.md for responsible disclosure guidelines.

---

❓ Troubleshooting

<details> <summary><strong>Agent service won't start</strong></summary>

1. Ensure you're running PowerShell as Administrator
2. Check .NET 8 is installed: dotnet --list-runtimes
3. Review Windows Event Viewer → Application log for WinSentinel.Agent errors
4. Try running the agent manually first: dotnet run --project src/WinSentinel.Agent

</details>

<details> <summary><strong>Dashboard can't connect to agent</strong></summary>

1. Verify the agent service is running: .\Install-Agent.ps1 -Status
2. Named pipe connections require both processes to run under the same user (or SYSTEM)
3. Some antivirus software blocks named pipe creation - add an exclusion for WinSentinel.Agent.exe

</details>

<details> <summary><strong>False positives in audit results</strong></summary>

1. Use Finding Rules in the dashboard to suppress known-acceptable findings
2. Switch to a more appropriate Compliance Profile (e.g., Home vs Enterprise)
3. Use --modules flag in CLI to skip irrelevant audit modules
4. Open an issue if a detection rule is genuinely wrong

</details>

<details> <summary><strong>Build fails on x86 or ARM</strong></summary>

WinSentinel targets x64 only. Always pass -p:Platform=x64:

dotnet build WinSentinel.sln -p:Platform=x64
dotnet test -p:Platform=x64

</details>

<details> <summary><strong>Network score seems wrong</strong></summary>

1. Run .\Fix-Network.ps1 to apply recommended network hardening
2. Some findings (LLMNR, NetBIOS) require registry changes + reboot
3. VPN adapters can trigger false "open port" findings - suppress with Finding Rules

</details>

---

🤝 Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Test your changes (dotnet test -p:Platform=x64)
4. Push and open a Pull Request

Ideas: plugin system for custom modules, Linux port, UI themes, localization, additional compliance profiles.

---

📄 License

MIT License - see LICENSE for details.

---

<div align="center">

Built with C# on .NET 8 · 72k+ LOC · 4,173 tests · Always watching 🛡️

⭐ Star · 🐛 Report Bug · 💡 Request Feature

</div>