Coding agents are great inside a single session — they inspect code, use tools, make edits, and explain their work. The trouble starts when a task is ambiguous, tied to specific exit criteria, long-running, or anchored in a large codebase. You end up reminding the agent of the process, copying output between sessions, second-guessing whether it followed the right steps, and deciding when a human needs to review the work. Atomic turns that process into code. A workflow can branch, retry, run stages in parallel, isolate sessions, pass only the right transcript forward, pause for human approval, and run inside a devcontainer so the agent is never loose on your host machine.Documentation Index
Fetch the complete documentation index at: https://bastani.mintlify.app/llms.txt
Use this file to discover all available pages before exploring further.
| Start with your own process | Automate the repetitive parts of research, debugging, review, migrations, or PR prep. One TypeScript file, versioned with the repo. |
| Scale to your team | Encode review gates, quality checks, and approvals so every teammate runs the same workflow instead of manually steering an agent. |
| Keep the coding agent | Atomic adds structure around Claude Code, OpenCode, and Copilot CLI without rebuilding file editing, tool use, MCP setup, hooks, or context handling. |
| Start in natural language | Ask the workflow-creator skill to turn a description into defineWorkflow() code — or let an agent invoke it when a complex task warrants a repeatable workflow. |
| Own the outer loop | Workflows, gates, handoffs, and the execution graph are TypeScript you can read, edit, and version — instead of a black-box harness improvising process. |
Where Atomic fits
Markdown is great for guidance: conventions, commands, repo notes, and checklists. Use Claude Code, OpenCode, or Copilot CLI directly for normal single-session coding. Reach for Atomic when the work needs branching, retries, parallel sessions, explicit state, human approval, sandboxed execution, or reliable handoff between stages. General agent frameworks can do some of this, but you often rebuild coding-agent basics yourself: file editing, terminal interaction, MCP setup, hooks, session handling, and repo-specific context. Atomic starts from production coding agents and adds the workflow layer around them.What you get
Atomic ships two surfaces that share one runtime — see CLI vs SDK:CLI
Pre-built autonomous behavior — context-file generation, the Ralph loop, deep research, and any registered workflow, run from your terminal.
SDK
@bastani/atomic-sdk — encode your team’s process as defineWorkflow() code and embed it in any TypeScript project or parent CLI.Next
Workflows
Stages, the execution graph, session isolation, transcript hand-offs, and human gates.
Quickstart
Generate context files, run Ralph, and build your first workflow.