-
-
Notifications
You must be signed in to change notification settings - Fork 80
Home
Vinh Nguyen edited this page Aug 7, 2026
·
3 revisions
Welcome to the VT Code wiki. VT Code is a Rust coding agent built for long-running autonomous workflows — with OS-native sandboxing, multi-provider LLM support, open protocols, and extensible Agent Skills.
Secure, open, universal.
VT Code is a terminal-based AI coding agent that:
- Runs an interactive TUI with slash commands, streaming, and session resume
- Supports 21+ LLM providers — Anthropic, OpenAI, Gemini, OpenRouter, plus local inference via Ollama, LM Studio, and llama.cpp
- Ships safe coding tools: file ops, ripgrep search, ast-grep structural search, fuzzy discovery, project indexing, and terminal execution
- Extends through Agent Skills, MCP client/server, lifecycle hooks, subagents, and ACP (Agent Client Protocol for Zed)
- Prioritizes safety: restricted shell sandbox, tool guardrails, subprocess isolation, dotfile protection, and full audit logging
- Supports open protocols: Open Responses, Agent2Agent (A2A), ATIF, and Anthropic Messages API
- Enables loop engineering: worktree isolation, propose/verify sub-agents, durable loop state, and cost guardrails
Install via Homebrew:
brew trust vinhnx/tap
brew install vinhnx/tap/vtcodeOr via the one-liner (installs ripgrep + ast-grep too):
curl -fsSL https://raw.githubusercontent.com/vinhnx/vtcode/main/scripts/install.sh | bashScaffold and launch:
vtcode init # creates vtcode.toml, .vtcode/, and AGENTS.md
vtcode # interactive TUICommon commands:
vtcode # interactive TUI
vtcode init # scaffold project config + AGENTS.md
vtcode ask "explain Rc vs Arc" # one-shot question (stdout is answer-only)
vtcode exec "refactor main.rs" # headless task with full tool access
vtcode review # review uncommitted changes
vtcode update # self-update| Page | What you'll learn |
|---|---|
| Getting-Started | Install, API keys, first session, onboarding |
| Interactive-Mode | TUI shortcuts, slash commands, fullscreen mode, background bash |
| Configuration |
vtcode.toml — agent, tools, commands, hooks, security |
| Providers | All supported LLM providers, custom providers, governance |
| Local-Models | Ollama, LM Studio, and llama.cpp local inference |
| Agent-Skills | The open Agent Skills format: discover, create, validate |
| Agent-Plugins | Portable skill + MCP packages: install, manage, configure |
| MCP-Integration | Connect external MCP servers, allowlists, embedded server |
| Full-Automation |
--full-auto, propose/verify, orchestrated harness, evals |
| Planning-Workflow | The /plan workflow and review gate |
| Security | Security model, sandboxing, tool policies, best practices |
| FAQ | Common questions and troubleshooting |
VT Code is built by an open source community. Contributions of all kinds are welcome — see the CONTRIBUTING guide and good first issues.
If VT Code has helped you ship something, consider a GitHub sponsorship. It means the world.
First-party code is licensed under MIT OR Apache-2.0. Third-party code remains under its original licenses.