One home for AI agents. One shared project brain.

AgentDock turns a GitHub repo into reusable context for Codex, Claude, Cursor, OpenClaw, Cline, and custom agents through CLI, generated files, and MCP.

Start building See workflow

Claude

Keeps product intent, project constraints, and decisions in view.

Codex

Uses the same repo brain to edit, test, and ship changes.

OpenClaw

Reads generated files and shared context without a reset.

Context anywhere

Now your agents can use the same context, anywhere you work.

AgentDock keeps your repository setup, product decisions, skills, memories, and handoffs in one clean workspace — then turns that context into files and workflows your agents can actually read.

Import a GitHub repo and create a project brain automatically.

Save the rules, memories, skills, and decisions your agents need.

Use the same context in your terminal, generated files, and MCP-ready workflows.

Import your repo Generate files

One project brain

Generated context for Claude, Codex, Cursor, Cline, OpenClaw, and MCP-ready agents.

How it works

From repo to agent-ready context in minutes.

AgentDock gives your AI agents the project context they usually miss: setup, decisions, skills, secrets, sessions, and generated files.

Create project from GitHub

Paste your repository URL. AgentDock fetches the repo name, description, tech stack, scripts, and setup details.

Store important memory

Save project rules, current bugs, product decisions, commands, and things every AI agent should remember.

Add skills for agents

Create reusable instructions like Supabase debugger, clean UI builder, security reviewer, or Codex prompt writer.

Save secret references safely

Do not expose real API keys. Store only safe references like agentdock://secrets/OPENAI_API_KEY.

Generate agent files

Create AGENTS.md, CLAUDE.md, Cursor rules, CLINE.md, OPENCLAW.md, and shared context files.

Create project brain Generate files

Access methods

Three clean ways to give agents the right context.

AgentDock does not force every agent into one setup. It creates the right context path for each tool — terminal, generated files, or MCP-ready workflows.