HubLensAI Agentsalibaba/anolisa
// archived 2026-04-27
alibaba

anolisa

AI🌱 NEW PROJECT BOOST#AI Agents#Operating System#eBPF#System Administration
View on GitHub
173

// summary

ANOLISA is an evolution of Anolis OS designed specifically to support AI agent workloads at the server-side operating system level. The project provides a comprehensive suite of components including an AI-powered terminal, security kernels, and observability tools. Users can easily integrate these features into their systems through standard RPM package installations.

// technical analysis

ANOLISA is an Agentic OS implementation designed to optimize server-side operating systems for AI Agent workloads by integrating specialized architectural layers. It addresses the unique challenges of AI-driven environments, such as security, observability, and token efficiency, by providing a modular suite of tools that bridge the gap between traditional OS functions and LLM-based automation. By leveraging technologies like eBPF and Qwen Code, the project makes a strategic trade-off to prioritize deep system integration and performance monitoring over generic, high-level agent frameworks.

// key highlights

01
Copilot Shell provides an AI-powered terminal assistant that facilitates code understanding and system management using Qwen Code.
02
Agent Sec Core implements an OS-level security kernel to handle sandboxing, system hardening, and asset integrity verification.
03
AgentSight utilizes eBPF technology to offer zero-intrusion observability into LLM API calls, token usage, and process behaviors.
04
Token-less offers a toolkit for LLM optimization, reducing operational costs through response compression and command rewriting.
05
OS Skills provides a curated library of pre-built functions for DevOps, cloud integration, and system administration tasks.

// use cases

01
AI-powered terminal assistance for system management and automation
02
OS-level security hardening and sandboxing for AI agents
03
eBPF-based observability for monitoring LLM API calls and token usage

// getting started

To begin using ANOLISA, install the full suite of components on your system using the command 'sudo yum install copilot-shell agent-sec-core agentsight tokenless os-skills'. Once installed, you can immediately start interacting with the AI-powered terminal by running the 'cosh' command.