How to use the DeepTutor CLI agent skill (deeptutor-cli)

What is an agent skill and what does this one do?

An agent skill is a folder with a SKILL.md file that gives an AI agent extra instructions and reference material for a specific task. When a request matches the skill's description, the agent reads those instructions and follows them.

Once installed, this skill tells the agent how to use the DeepTutor command-line interface. The agent can install and configure the platform, check whether it is ready, create and search knowledge bases, start quizzes, deep solve and deep research runs, and launch the API server or Web app. The skill asks the agent to read the project's setup guide first, reuse existing checkouts, refer to API keys by environment variable instead of printing them, and report only verified readiness.

Install the deeptutor-cli skill

Download the skill ZIP from this page and unzip it. You can install it in one of two ways.

For Claude Code, place the unzipped folder in ~/.claude/skills/ to make it available for all your projects, or in a project's .claude/skills/ folder to limit it to that project.

In the Claude apps, open Settings, go to Capabilities, then Skills, and upload the ZIP.

• Personal (all projects): ~/.claude/skills/
• Single project: .claude/skills/ inside the project folder
• Claude apps: Settings → Capabilities → Skills, then upload the ZIP

Prepare DeepTutor before you ask for setup

The skill assumes you have Python 3.11 to 3.14 and pip. Node.js 22 LTS with npm is optional and only needed for the Web UI. You also need an API key from an LLM provider, which you supply through an environment variable. Optional embedding and search provider credentials are needed for RAG and web search.

The skill's setup guide describes a first-time interactive setup with deeptutor init. For CLI-only use, add --cli to skip the ports step. Use --home <path> to point at a specific workspace. Setup cannot finish without a model provider key, so have one ready.

• Python 3.11–3.14 and pip
• An LLM provider API key set as an environment variable
• Optional: Node.js 22 LTS and npm for the Web UI
• Optional: embedding and search provider credentials for RAG and web search

Trigger the skill with natural requests

You do not need to type a command name. Describe the outcome you want in plain language and mention DeepTutor. The agent matches your request to the skill's description and chooses commands from its reference.

Requests that fit the skill include setting up or diagnosing DeepTutor, building a knowledge base from your files, running a quiz or deep research task, and starting the Web app.

• "Set up DeepTutor on this machine and check that it is ready."
• "Diagnose why my DeepTutor install won't start."
• "Create a knowledge base called biology-notes from the PDFs in ./biology."
• "Generate five quiz questions on linear algebra with DeepTutor."
• "Start the DeepTutor API server so I can test it."

Run DeepTutor capabilities from the terminal

The skill's command reference includes one-shot runs and an interactive chat. Capabilities such as deep_solve, deep_question (quiz generation), deep_research and visualize are started with deeptutor run. The skill notes that core installs provide chat and ask_questions, while other capabilities require their owning modules. Run deeptutor plugin list to see what is registered in your installation.

Knowledge base commands cover listing, creating from documents or a directory, adding documents, searching, and evaluating retrieval against a JSONL question set.

• Use --kb <name> to mount a knowledge base for RAG
• Use --format json for machine-readable output
• Use deeptutor kb search <name> "query text" to test retrieval directly

Safety notes: scripts and network access

This skill runs scripts and uses the network. Setup and online checks send requests to model providers, and knowledge base creation can send document content to the providers you configure. The skill's own guidance is to reference credentials by environment variable and never print or write literal keys into setup files.

Before uploading sensitive documents, check which LLM, embedding and search providers you have configured. If they are cloud services, the content may leave your machine. Review the setup JSON and the commands the agent proposes before approving them, especially any that change settings or start services.

• deeptutor doctor --online makes a small live request to your LLM provider
• Use --check to validate a setup file without writing settings or using the network
• Stop DeepTutor before changing modules, as the skill instructs

Tips for getting reliable results

Ask the agent to run deeptutor doctor first so you know the current state. Saving settings is not the same as being ready, so ask for the verification output rather than a yes or no. If you are working with an older release, ask the agent to check --help output, since the skill notes that command names can differ between versions.

Keep DEEPTUTOR_HOME the same across commands so the agent does not split your data between two workspaces. Point the agent at an existing checkout rather than a fresh clone if you already have one.

• Start with a readiness check, then make changes
• Name the workspace with --home or DEEPTUTOR_HOME
• Ask the agent to report what it verified and what remains unfinished