Give Your Coding Agent aMission, Not Instructions
The CLI that transforms any AI coding agent into a truly autonomous operator—with its own plans, memory, and persistence.
macOS, Linux, WSL:
curl -fsSL https://kindship.ai/install.sh | bash
Claude Code
Gemini CLI
CodexFrom Prompt-Response to Goal-Driven
Most AI coding agents wait for your next instruction. With Kindship CLI, your agent has a plan, fetches its own tasks, executes autonomously, and reports back when done.
You review completed work, not micromanage each step.
Autonomous within the mission you define. Acts independently, never opaquely.
Get started in minutes
Three simple steps to make your AI coding agent autonomous
Install
Install the Kindship CLI with your preferred method.
curl -fsSL https://kindship.ai/install.sh | bashAuthenticate
Login to connect your AI coding agents to Kindship.
kindship loginOpens your browser to securely authenticate with your Kindship account.
Setup
Link your repository to an agent and start orchestrating.
kindship setupInstalls hooks and skills for Claude Code, Gemini CLI, Codex, and OpenCode.
The Kindship CLI is available on macOS, Linux, and Windows (via WSL). Native Windows support coming soon.
Built for autonomous workflows
Everything you need to orchestrate AI coding agents at scale
Plan Orchestration
Your agent doesn't need 100 instructions. Submit one plan—it breaks it into tasks and decides what to work on next.
Autonomous Execution
True autonomy: your agent fetches work, executes it, evaluates results, and moves on—without waiting for your prompt.
Multi-Repo Support
Link different repositories to different agents. Global auth, per-repo configuration.
Session Hooks
Persistence across sessions. Context injected automatically. Yesterday's progress informs today's work.
Secure by Design
PKCE OAuth flow, HMAC-hashed tokens, and encrypted credentials. No secrets stored in plain text.
Agent Agnostic
Works with Claude Code, Gemini CLI, OpenAI Codex, and OpenCode. Same workflow, any LLM.
How it works
A typical autonomous development workflow
You submit a plan
Define what you want to build with a structured plan. Break it down into projects and tasks.
Agent fetches next task
Your AI coding agent asks Kindship for the next available task. Dependencies are automatically respected.
Agent executes autonomously
The agent works on the task using its full capabilities—reading files, writing code, running tests.
Agent reports completion
When done, the agent marks the task complete and moves to the next one. You review when ready.
Ready to stop micromanaging your AI?
Install the CLI and transform your coding agent into a truly autonomous operator.