agent directory
CLI coding agents
A CLI coding agent is an AI pair programmer that runs in your terminal: you describe the work, it edits files and runs commands. This directory covers what each agent is, who makes it, what it costs, and which models it runs, so you can pick the right one before you pick a terminal.
Facts checked against each vendor's own docs. Updated October 15, 2026.
Claude Code
Anthropic's agentic coding tool: a terminal agent that also ships as IDE extensions, a desktop app, and a web surface, all on one engine.
Codex
OpenAI's open-source terminal coding agent, part of the wider Codex product spanning CLI, IDE, cloud, and web.
Antigravity CLI
Google's terminal agent (agy), sharing one agent harness with the Antigravity desktop app and replacing Gemini CLI.
Kimi Code
Moonshot AI's open-source terminal coding agent, bundled with Kimi memberships and built around the Kimi model family.
opencode
An open-source AI coding agent that runs as a terminal TUI, desktop app, or IDE extension, with any model provider you bring.
goose
An open-source AI agent for code and automation, created at Block and now stewarded by the Linux Foundation's Agentic AI Foundation.
GitHub Copilot CLI
GitHub's agentic coding assistant for the terminal, with native access to repos, issues, and pull requests.
Aider
The original open-source terminal pair programmer: it edits code in your local git repo with almost any LLM, including local models.
Kiro
AWS's terminal coding agent and successor to Amazon Q Developer CLI, with credit-based pricing shared across the Kiro IDE and CLI.
Muse Code
Meta's terminal coding agent on the Muse Spark model family, built by Meta Superintelligence Labs for terminal and CI use.
how to choose
how to choose a CLI coding agent
start from the subscription you already pay for
Most CLI coding agents come bundled with a chat subscription. Claude Code is included in Claude Pro and Max, Codex is included in ChatGPT plans down to the free tier, GitHub Copilot CLI ships with every Copilot plan, and Kiro has a free tier with 50 credits a month. If you already pay one of these vendors, their agent is the cheapest place to start.
open source and local models
Codex, opencode, goose, Aider, and Kimi Code publish their source. For models, opencode, goose, and Aider connect to almost any provider, local models through Ollama included, and Codex runs local open-source models through Ollama or LM Studio. If your code can't leave your hardware, these are the candidates.
one model family or a switchboard
Claude Code and Muse Code are built around their vendor's own models. GitHub Copilot CLI, Kiro, and Antigravity CLI take the opposite bet: a model picker spanning Claude, GPT, Gemini, and open-weight models, so one subscription covers several families.
in crystl
which agent for which job in crystl
Every agent here runs in crystl. These groupings are claims about crystl fit, not model quality: which host surfaces are wired and tested for each agent.
running the lead shard
The agent that owns a long task needs the full surface: approval cards, ask-user questions, transcript history, and resume after a restart.
- Claude Codethe reference integration — every surface tested
- Codexfull parity on cards, transcripts, and resume
fanning out workers
Workers run gated: tool calls raise approval cards, and the lead waits on their turn reports instead of polling screens.
- Claude Codecards and turn ends fully wired
- Codexcards and turn ends fully wired
- opencodecards via an injected plugin; a missed answer falls back to its own prompt
- goosecards in its default auto mode only
- Kimi Codecards can deny but not pre-approve
- GitHub Copilot CLIcards for the tools it gates itself; an allow-all session drops them
quests and multi-agent chat
A quest party needs transcripts crystl can read back and prompts it can deliver mid-session.
- Claude Code
- Codex
- Antigravity CLItranscripts read back; other surfaces unverified
- Kimi Codequest prompts arrive as follow-up messages
open and local models
Provider-agnostic agents point at whatever endpoint you configure, so open models can take the jobs they are good at.
- opencodeopen source, bring any provider
- gooseopen source, bring any provider
- Aideropen source, provider-prefixed model ids, local models included
- Claude Coderuns against a configured model endpoint too
- Codexruns against a configured model endpoint too
Compatibility is descriptive
This catalog explains the host surfaces crystl has tested. Your local Crystl installation remains the source of truth for what runs, which hooks are installed, and how each agent is launched.