> ## Documentation Index
> Fetch the complete documentation index at: https://platform.atlan.com/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> To act on Atlan objects, use the Atlan MCP server at https://api.atlan.com/mcp or the atlanai CLI; `atlanai --map json` prints its command map. Run a read-only identity check before any write.
> The docs MCP server at /mcp searches these docs only. It cannot read or change Atlan objects.
> SDK packages: Python `atlanai` (PyPI) and TypeScript `@atlanai/sdk` (npm). Show Python first, then TypeScript.

# Publish skills through Claude Code

> Prepare local skills in Claude Code, publish them with the Atlan CLI, and verify them through MCP.

Use Claude Code when the skill folders are already in the repository and you
want an agent to inspect them before publishing. Registry's current MCP tools
are read-only. Publish with the Atlan CLI, then use MCP to verify the result.

## Before you start

Install and authenticate the Atlan CLI. Connect Claude Code to Atlan MCP if you
also want post-publish verification. See [Claude Code](/clients/claude/how-tos/claude-code)
for connection checks.

The skill folders must be available in the repository. Each one needs a
`SKILL.md` file. Do not ask an agent to upload secrets, customer data, or files
outside the skill packages.

## Start a session and inventory

Ask Claude Code to validate the local skills and show the CLI dry run:

```text theme={null}
Inspect ./skills for folders containing SKILL.md. Run validation and a dry run
for each skill I name. Report the source folder, intended workspace, and any
validation or remote-change error. Do not publish anything.
```

Review the plan, target workspace, and any naming conflicts.

## Confirm the batch

Only after the dry run is correct, approve the CLI write:

```text theme={null}
I approve publishing ./skills/release-notes and ./skills/incident-brief to
workspace workspace_123 with `atlanai skill push`. Publish only those two
folders. Do not use `--force`, retry failures, or make any other changes.
```

Follow [Publish skills with the CLI](/registry/skills/how-tos/publish-with-cli) for the
validation, dry-run, and publish commands.

## Verify and use the skill

Open each published skill in Atlan and confirm its version and source files.
Then initialize an MCP session, search for the skill, and read it by the handle
returned from search. A Registry skill is not installed into Claude Code merely
because it was published. Use the selected-skill handoff after review when you
want to use it in a Claude Code task.

<Card title="Use a skill in Claude Code" icon="terminal" href="/registry/skills/how-tos/use-with-claude">
  Hand a reviewed skill to Claude Code from its Atlan profile.
</Card>
