kavach

Tactical AI Workspace Monitor & EDR

253
Stars
43
Forks
TypeScript
语言
2026/8/23
添加时间

⚠️ 第三方软件声明

本 Skill 为第三方开源软件,独立托管于 GitHub。SkillTip 仅为信息目录,不控制或维护底层仓库。所显示的安全检查为自动化且范围有限,安装前请自行审查源码。

阅读服务条款

安装

添加到你的 Claude Code skills 目录:

# Add to your Claude Code skills
git clone https://github.com/LucidAkshay/kavach

快速入门

使用 kavach 等 Skills 的指南。

安全报告

已验证

上次扫描:—

{
  "status": "PASSED",
  "issues": []
}

README.md

🛡️ KAVACH

Tactical AI Workspace Monitor & EDR

Tauri Rust React License


Kavach | Product Hunt



Kavach Dashboard Demo


Kavach (Sanskrit for "Armor") is an open source Endpoint Detection and Response (EDR) layer designed exclusively to restrain, monitor, and remediate autonomous AI agents and local LLMs operating on your machine.


🛰️ The Threat & The Observer

Autonomous AI agents are revolutionary, but they are inherently volatile. When they operate on your local file system, they move at superhuman speeds. A single hallucination can delete production databases, overwrite source code, or exfiltrate private API keys.

Architecture Transparency: Kavach v1.1.0 operates as a high performance asynchronous file system observer using Userland APIs. It monitors designated workspaces, quarantines malicious actions, and provides immediate remediation tools. It is an EDR monitor, not a strict synchronous sandbox. True system call blocking requires kernel drivers, which are slated for our v1.2 roadmap.


⚡ The Arsenal : Active Defense Mechanisms

Kavach is armed with an exhaustive array of active defense mechanisms operating at the application layer to outsmart and contain rogue agents.

🌌 Flagship Tactical Mechanics

  • Phantom Workspace: Intercepts destructive file operations (Modify, Delete, Create) and silently routes them to a hidden .kavach_phantom directory to trap malicious payloads.
  • The Auto Enforcer: A ruthless background Rust thread. If an intercepted action sits pending in the UI for more than 60 seconds, Kavach automatically issues a hard OS termination signal to the rogue agent PID.
  • Temporal Rollback: Automatic cryptographic caching of file modifications. If a file is under 50MB, Kavach enables instant 1 click restoration. Large files are dynamically flagged in the UI as ineligible for rollback.
  • Honeypot Tripwire Architecture: Safely deploys a "sugar trap" system_auth_tokens.json file. Any process attempting to read or modify the decoy instantly triggers a High Risk lockdown and a distinct TRAP TRIGGERED alarm. Decoys are automatically scrubbed when the session ends.

🦀 Core Security Engine (Rust Layer)

  • High Velocity Loop Break: Heuristic detection of repetitive command patterns (e.g. recursive npm install loops) to force automatic agent suspension.
  • The Janitor Protocol: Intelligent auto whitelisting of standard Windows system maintenance paths (%TEMP%, Prefetch, Logs) to eliminate alert fatigue.
  • PID Chokehold: Dynamic monitoring and CPU usage throttling. Instantly drops a rogue agent's OS priority to IDLE using Windows WMIC or POSIX renice commands.

🧠 The Brain & The Shield

  • Gag Order (PII Sanitizer): Real time entropy and regex scanning for OpenAI AWS keys, credit cards, and emails in outbound strings.
  • Faraday Clipboard Guard: Actively calculates Shannon Entropy on your OS clipboard. If it detects a stolen secret, it instantly overwrites your clipboard with a decoy string to prevent pasting and exfiltration.
  • Child Process Quarantine: Hierarchical scanning that tracks and restricts permissions for any process spawned by a monitored parent agent.

📦 The Black Box (Forensics)

  • Cryptographic Ledger: A blockchain style FNV hash chain that ensures audit logs are immutable and tamper proof.
  • Supply Chain Auditor: Real time CVE scanning against the workspace package.json for known malware dependencies.
  • Predictive Blast Radius: Recursive import scanner that visually maps exactly what will break in your project before you approve a file deletion.

🚀 v1.2 Roadmap : True Sandboxing & Autonomous Defense

Moving from a passive EDR tool to a true Zero Trust Sandbox requires native hardware drivers and autonomous intelligence. These are currently under active development:

  • Autonomous Reasoning Loop (Local LLM): A lightweight, internally hosted reasoning agent that will automatically triage events, correlate velocity spikes with honeypot triggers, and dynamically rotate decoys without human intervention.
  • Linux eBPF Probes: Kernel level hooks to intercept sys_enter_openat and sys_enter_mkdir.
  • Windows Minifilters: Native file system drivers for absolute Windows execution blocking.
  • macOS Endpoint Security Framework (ESF): Strict entitlements for Apple silicon.
  • Network Ghost Mode: Local root certificate authority for intercepting and spoofing TLS traffic without payload execution.

⚙️ Under the Hood

Kavach is built for absolute performance and zero latency, running entirely locally on your machine with zero cloud dependencies.

  • The Engine (Rust): The core interception logic, OS execution hooks, and cryptographic file caching are written in Rust for memory safety and bare metal speed.
  • The Command Center (React TypeScript): The FUI dashboard operates in an isolated webview, ensuring the UI thread never blocks the security engine.
  • The Bridge (Tauri): Secure asynchronous Inter Process Communication (IPC) bridges the frontend dashboard and the Rust watcher.

💻 Zero Config Deployment

  1. Download: Grab the latest installer for your operating system from the Releases Page.
  2. Launch: Run the executable (e.g. Akshays.Kavach_1.1.0_x64-setup.exe). (Note: Windows requires running as Administrator for process termination; macOS requires Full Disk Access).
  3. Arm: Select your workspace directory in the UI. Kavach immediately locks down the perimeter.

📞 Direct Line & Feedback

Kavach features a "Direct Line" communication module for users to report anomalies, suggest features, or collaborate.

  • Developer: Akshay Sharma
  • Feedback: Use the [ COMM LINK ] terminal button inside the application to route encrypted feedback directly to the developer inbox via your OS default mail client.

📜 License

Kavach is proudly released under the GNU General Public License v3.0 (GPLv3). Permanent attribution to Akshay Sharma is required in all forks, distributions, and derivatives, and any modifications must remain open source under the same license terms.

常见问题

What is kavach?

kavach is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by LucidAkshay. Tactical AI Workspace Monitor & EDR. It has 253 GitHub stars.

Is kavach safe to use?

kavach returned warnings in SkillsLLM's automated security scan. It has no critical vulnerabilities, but review the flagged issues in the Security Report section before adding it to your workflow.

How do I install kavach?

Clone the repository with "git clone https://github.com/LucidAkshay/kavach" and add it to your Claude Code skills directory (see the Installation section above).

What programming language is kavach written in?

kavach is primarily written in TypeScript. It is open-source under LucidAkshay on GitHub, so you can review or fork the full source.

Are there alternatives to kavach?

Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh kavach against similar tools.

评论 (0)

暂无评论,成为第一个分享想法的人!

ECC

by affaan-m

10

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

242,21936,702JavaScript
AI 智能体ai-agentsanthropicclaude-code
查看详情
15

An agentic skills framework & software development methodology that works.

234,96620,863Shell
AI 智能体ai-agentsbrainstorming
查看详情

hermes-agent

by NousResearch

10

The agent that grows with you

234,43747,175Python
AI 智能体ai-agentsagent-orchestration
查看详情

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

185,94028,768JavaScript
AI 智能体ai-agentsanthropicclaude-code
查看详情

cc-switch

by farion1231

3

A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build & Hermes Agent. Only official website: ccswitch.io

128,8688,826Rust
AI 智能体claude-codeai-tools
查看详情

claude-code

by anthropics

Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.

120,03119,897Shell
AI 智能体
查看详情

开发者还喜欢

基于喜欢此 Skill 的开发者投票和收藏

ECC

by affaan-m

10

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

242,21936,702JavaScript
AI 智能体ai-agentsanthropicclaude-code
查看详情
15

An agentic skills framework & software development methodology that works.

234,96620,863Shell
AI 智能体ai-agentsbrainstorming
查看详情

hermes-agent

by NousResearch

10

The agent that grows with you

234,43747,175Python
AI 智能体ai-agentsagent-orchestration
查看详情

n8n

by n8n-io

12

Fair-code workflow automation platform with native AI capabilities. Combine visual building with custom code, self-host or cloud, 400+ integrations.

201,88160,308TypeScript
MCP 服务器apisai-tools
查看详情

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

185,94028,768JavaScript
AI 智能体ai-agentsanthropicclaude-code
查看详情

cc-switch

by farion1231

3

A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build & Hermes Agent. Only official website: ccswitch.io

128,8688,826Rust
AI 智能体claude-codeai-tools
查看详情