# Import skills in bulk

Bring existing skill files into Studio, choose what happens when one already exists, and read which files were skipped and why.

Skills written elsewhere, exported from another tenant, or kept in a repository come in as files. The importer takes a folder of them at once and tells you what it did with each one.

**Who can use it:** the **Builder** role or higher.

## What a skill file is

Markdown with YAML front matter: the front matter carries the name, version, description, vertical, role, tags and author, and the body is the instructions. It is the format **Export** produces, and it follows the public skill-file specification, so files written for that spec import unchanged.

Front matter the platform does not recognise is shown as extra fields and dropped on import rather than stored.

## Import a folder

![The Import skills screen: instructions to drop a folder of .md files, the drop zone with Or click to browse and Pick a folder instead.](https://docs.katonic.ai/img/agents/skills/skills-import.png)
*Import takes SKILL.md files in bulk. Files without valid front matter and a body are skipped, and the screen says why.*

  
**Step 1. Give it the files**

Drag Markdown files or a folder onto the drop area, or select it to browse. **Pick a folder instead** opens a folder chooser; the browser asks you to confirm access to the folder, which is the browser asking rather than the platform. The two are separate on purpose, so choosing a single file never triggers the folder prompt.

  
**Step 2. Read the list before importing**

Every file is parsed as you add it and listed either as a name and version with its description, or as a filename and the reason it cannot be imported. The drop area counts them: how many are valid, how many are not, how many there are.

  
**Step 3. Choose what happens to duplicates**

**On conflict** decides what to do when a skill of that name and version already exists: **Bump**, the default, saves the file as a new version named after the old one with an import suffix; **Skip** keeps what you have; and **Error** stops the import.

**Default vertical** and **Default role** apply only to files whose own front matter leaves them out, so a folder of files from one team can be classified in one go.

  
**Step 4. Import**

The button counts what it will do. Afterwards, the result says how many of the total were imported, and marks each file as created, bumped, skipped or failed. **Import more** returns to the drop area; **View library** opens the library.

## Why a file was skipped

| The reason | What it means |
| --- | --- |
| No YAML front matter | The file does not begin with a front-matter block |
| Front matter not closed | The block was opened and never closed |
| Missing a required field | The front matter has no name, or no description |
| The body is empty | There is front matter but no instructions |

Nothing invalid stops the rest: the valid files still import, and the invalid ones are listed with their reason.

## Troubleshooting

- **Dragging a folder does nothing.** Not every browser supports folder drops. The importer says so when it detects one, and the file picker handles folders in every browser.
- **A skill imported under a name you did not expect.** The name comes from the front matter, not the file name.
- **Everything was skipped.** Check the conflict setting. **Skip** keeps what you already have, so re-importing the same files under it changes nothing; the default is **Bump**, which saves them as new versions instead.

## Related

- [Skills](/agents/skills)
- [Write a skill](/agents/skills-author)
- [Attach a skill to an agent](/agents/skills-attach)

Source: https://docs.katonic.ai/agents/skills-import
