Documentation

docs/getting-started.md

Getting started

Install AgentKrew into an existing project, verify it, and run your first multi-agent workflow. The whole flow takes a few minutes.

docs/agents.md

Agents

An agent is a specialized role in your AI team: a named expert with a clear responsibility, a description that tells the runtime when to delegate to it, and a prompt that defines h…

docs/skills.md

Skills

A skill is reusable engineering knowledge: guidance a loaded agent applies on demand — PostgreSQL schema rules, systematic debugging, TDD discipline. Unlike agents, a skill has no …

docs/commands.md

Commands

A command is a user-facing entry point: the thing you type to kick off specialized work. Commands contain no agent logic of their own — each command names exactly one target, eithe…

docs/workflows.md

Workflows

A workflow is the orchestration layer: an ordered, declarative plan that routes one request through several agents — some in sequence, some in parallel — and defines what "done" me…

docs/customization.md

Customization

AgentKrew's update system is built on one rule:

docs/updates.md

Updates

agentkrew update refreshes an existing installation to a newer AgentKrew version — and the whole design exists to answer one question: will I lose my customizations? The answer is …

docs/troubleshooting.md

Troubleshooting

Almost every AgentKrew problem begins and ends with:

docs/runtimes/claude.md

Claude Code runtime

Claude Code is AgentKrew's default runtime. The Claude adapter renders canonical artifacts into Claude-native project files — no settings or CLAUDE.md edits required.

docs/runtimes/codex.md

Codex runtime

The Codex adapter renders canonical artifacts into Codex-native project files: agent TOML files under .codex/ and skills under .agents/.

docs/runtimes/opencode.md

OpenCode runtime

The OpenCode adapter renders canonical artifacts into OpenCode-native project files under .opencode/ — agents, commands, and skills as plain markdown, with no opencode.json edits r…