Witty Assistant CLI Agent Introduction ​

Introduction ​

This manual provides a comprehensive introduction to the Witty Assistant CLI agent capability system. The agent series is an intelligent interaction tool built on the Witty Assistant platform for vertical business scenarios. Relying on a dedicated technical architecture and MCP service capability foundation, it deeply adapts to business scenario requirements, enabling lightweight, scenario-specific intelligent service deployment.

Currently, the Witty Assistant platform has integrated two agents: LLM chat and Basic Operations Agent. LLM chat focuses on intelligent dialogue inquiry capabilities, providing users with natural language interaction support. The Basic Operations Agent is a specialized tool for openEuler basic operations, with its core advantage lying in the integration of two major MCP service capabilities, achieving an organic combination of full-scenario openEuler system basic operations coverage and lightweight knowledge retrieval. The Known Issues Analysis Agent will be developed for known issue diagnosis scenarios. Through standardized capability descriptions and practical use cases, this manual provides enterprise operations personnel with "ready-to-use" operational guidance, helping to lower the operational threshold and enhance the standardization and efficiency of operations work.

Default Agent Summary Table ​

LLM chat ​

LLM chat is an intelligent dialogue agent integrated into the Witty Assistant platform. It relies on the capabilities of mainstream LLM models and does not require MCP service support. It focuses on providing users with efficient, natural intelligent dialogue inquiry services and serves as the foundational module of the Witty Assistant platform's intelligent interaction system.

Core Capabilities ​

LLM chat leverages mainstream LLM models configured by users, possessing efficient and accurate natural language conversation and inquiry capabilities without requiring MCP services or complex configurations. Positioned as a "lightweight intelligent conversation assistant," it has no professional operational threshold. Users can directly initiate inquiries through natural language and quickly receive clear, accurate responses and feedback.

Use Case ​

The following demonstrates a basic use case:

Introduce openEuler

Basic Operations Agent ​

The Basic Operations Agent integrates two core MCP services to create a one-stop operations hub combining "basic operations + knowledge retrieval." All tools are adapted to openEuler system characteristics, support local/remote execution (for some tools), provide bilingual (Chinese/English) returns, and feature strict parameter specifications with unified return formats, allowing direct integration into enterprise operations processes.

Core Capabilities ​

Use Cases ​

The following demonstrates file operations, providing natural language interaction prompt formats with key information such as file paths for immediate use. All scenarios are tailored to actual enterprise operations needs.

• Scenario: Local File Batch Management

  text

    Need to manage files in the local /root/test_python directory, perform the following operations:
       1. View detailed information (including permissions, size, modification time) of all files in the /root/test_python directory;
       2. Create an empty folder named backup_202601 in the /root/test_python directory;
       3. Move all files with .py extension from the /root/test_python directory to the backup_202601 folder;
       4. Verify the move results and return the list of successfully moved files.

MCP Overview ​

The following details each Server's information and the core details of their subordinate tools.

MCP_Server List ​

MCP_Server Details ​

oe-cli-mcp-server ​

| oe-cli-mcp-server | pkg_tool | Supports openEuler system software package query, install, update, remove, cache cleanup, based on dnf/rpm with secure invocation | - action: Operation type (required, enum: list/info/install/local-install/update/update-sec/remove/clean)
- Other parameters: pkg_name (package name), rpm_path (absolute RPM path), cache_type (cache type), etc. (passed as needed) | success (execution result), message (execution info/error message), result (operation results list/log), pkg_name (operated package name), target (fixed as 127.0.0.1) | | | proc_tool | Supports process/service view, start, stop, force termination management operations, based on ps/systemctl/kill commands | - proc_actions: List of operation types (required, enum: list/find/stat/start/restart/stop/kill)
- Other parameters: proc_name (process name), pid (process ID), service_name (service name), etc. (passed as needed) | success (operation result), message (operation info/error message), result (operation results list/log), target (fixed as 127.0.0.1), proc_actions (executed operations) | | | ssh_fix_tool | Diagnoses and fixes SSH connection failures on openEuler systems, checks port connectivity, sshd service status, and repairs configuration | - target: Target host IP/hostname (required)
- port: SSH port (default 22)
- lang: Language setting (optional) | success (fix result), message (fix result description), target (target host), result (list of execution results for each step) | | | sys_info_tool | Batch collects system, hardware, and security information from openEuler systems, pure Python + safe system command invocation | - info_types: List of information types (required, enum: os/load/uptime/cpu/mem/disk/gpu/net/selinux/firewall) | success (collection result), message (collection info/error message), result (structured collected data), target (fixed as 127.0.0.1), info_types (collected types) |

rag-server ​