A Claude Code skill that teaches Claude to author workflows: deterministic multi-agent orchestration scripts that fan work out to fresh-context subagents under plain JavaScript control flow.
A workflow is a JavaScript file. The loops, the conditionals, the fan-out are
ordinary code that you control. Only the leaf agent() calls spend model tokens,
and each one runs in its own clean context window. The result is multi-agent work
that behaves the same way every run and can be resumed if it stops partway.
This skill carries the file format, the judgement calls, and a tested authoring procedure, so you can just ask Claude to "create a workflow for X" and get a correct, runnable file back.
This skill writes files for Claude Code's Workflow tool. As of right now that tool is unreleased and unannounced.
It ships inside the Claude Code binary, but it stays hidden behind an environment variable, and Anthropic has not officially documented or announced it. That can change at any time, including how it is enabled and how it behaves.
What this means for you:
- You can install this skill today and read through it to understand workflows.
- The workflows it writes will only actually run once the Workflow tool is switched on in a released version of Claude Code.
- The right move is to wait for Anthropic to officially announce and release the feature, then come back and use this skill once it is live.
Treat this repo as a preview until then. When the feature ships, everything here is ready to go.
The repo root is the skill itself. Drop it into your Claude Code skills folder and Claude picks it up automatically.
| Path | What it is |
|---|---|
SKILL.md |
The skill entry point: the procedure Claude follows to design and write a workflow. |
references/api-reference.md |
The complete manual: every global, every option, every cap and constant. |
references/patterns.md |
Copy-paste orchestration patterns (fan-out, pipeline, loop-until-budget, judge panel, and more). |
assets/templates/ |
Starter files for the three core shapes: fan-out, pipeline, loop. |
assets/examples/ |
Six complete, runnable example workflows, with a README mapping each one to a technique. |
scripts/validate-workflow.mjs |
A linter that checks a workflow file against the parser's hard rules before you run it. |
Copy the skill into your global Claude Code skills folder:
git clone https://github.com/ray-amjad/claude-code-workflow-creator.git
mkdir -p ~/.claude/skills
cp -R claude-code-workflow-creator ~/.claude/skills/workflow-creatorThat is all. The next time Claude Code starts, the skill is available. Ask Claude to "create a workflow" and it will trigger.
-
Enable the Workflow tool. Today it is gated behind an environment variable:
export CLAUDE_CODE_WORKFLOWS=1 claudeWhen the feature is officially released, follow Anthropic's instructions for turning it on, since the exact mechanism may change.
-
Ask Claude to build one, for example: "create a workflow that reviews my branch across bugs, security, and tests, then verifies each finding."
-
Claude uses this skill to write the file, validate it, and run it. Watch live progress with the
/workflowscommand.
The details in this skill (the file format, the determinism rules, the caps, how to set a model, how structured output works) were checked directly against the Claude Code binary, not guessed. Because the feature is pre-release, some of it may shift before launch. If something stops matching once the feature is officially out, open an issue.
Built and maintained by Ray Amjad. If this is useful, the channel covers more agentic coding workflows like this one.