Skip to content

AI Assistant

OpsMini has built-in AI large-model operations capabilities, using natural language to perform system diagnostics, log analysis, and command execution.

Capabilities

  • Natural-language diagnostics: describe the fault symptoms, and the AI gives diagnostic suggestions based on real-time system status
  • Log analysis: let the AI analyze abnormal logs and locate the root cause of errors
  • Command execution: the AI converts natural-language intent into operations commands, which are executed after permission confirmation
  • Deep thinking: optional "think first, then answer" mode for complex reasoning problems
  • Built-in tools: injects host system information in real time (CPU/memory/disk/processes/ports/firewall), and supports searching for and installing skills in SkillHub via function calls

Supported LLM Providers

Provider Description
openai OpenAI-compatible interface
deepseek DeepSeek
qwen Qwen (DashScope-compatible interface)
ollama Local Ollama deployment

Configuration

Configure everything in Panel "Settings → AI Model Integration" (the original ai section of config.yaml has been migrated here and takes effect immediately at runtime):

  • Enable AI assistant (master switch)
  • Model provider / model name / API Key / API address
  • Natural-language execution / intelligent log analysis / intelligent alert diagnosis toggles

Configuration priority: environment variable OPSMINI_AI_KEY > panel settings > residual ai.* in the configuration file. It is recommended to store the key in an environment variable to avoid writing it to disk. See Panel Settings for details.

Security Mechanisms

  • When the AI determines "execution intent", it generates a structured action + parameters
  • Execution happens only if it hits the permission allowlist; unauthorized / high-risk operations require a second user confirmation
  • Executed actions are recorded in the audit log
  • After turning off the "Enable AI assistant" switch, AI-related interfaces are rejected outright

MCP Management

Supports configuring MCP (Model Context Protocol) services to extend the AI assistant's capabilities and data sources. You can create, read, update, and delete MCP configurations in "AI Assistant → MCP", supporting both command (stdio) and URL (SSE/HTTP) transport methods.

Skills (Skill Hub)

A built-in "Skills" tab connects to the SkillHub skill marketplace:

  • Search and one-click install skills
  • Upload local skills and manually create skills
  • Enable/disable and delete installed skills