Usar LetronnaLa app
Skills and your own commands
Saved instructions Letronna can follow, and prompts you run as slash commands.
Esta página aún no está traducida, así que se muestra en inglés. Puedes traducirla en GitHub. GitHub ↗
A skill is a set of instructions Letronna can follow when a task needs it, with any files those instructions use. Letronna sees each skill's name and description, and reads the rest only when it uses one.
Where skills come from
- Built in: a few come with Letronna, including one that explains Letronna's own commands and settings.
- Yours: in the
skillsfolder of your data folder. - The project's: in
.agents/skillsor.claude/skillsin the project folder. - Learned: skills the model saves from work that went well. They start switched off until you switch them on.
- From plugins: see Plugins, hooks and more.
When two skills have the same name, yours wins.
Writing one
A skill is a folder with a SKILL.md in it: frontmatter with its name and description, then the instructions. It is the same layout Claude's skills use, so those work here too.
---
name: release-notes
description: Write release notes from the git log since the last tag.
---
1. Find the last tag with `git describe --tags --abbrev=0`.
2. ...In the window, open Capabilities and choose New skill to write one there. A built-in skill can be copied with Copy to edit.
Adding and switching skills
In the terminal:
/skills
/skills add owner/repo
/skills on release-notes
/skills off release-notes/skills lists every skill and uses one; /skills add adds skills from a GitHub repository, a link or a folder. In the window, the same is on the Capabilities page.
Every skill that is switched on also runs as a slash command by its name, such as /release-notes.
Your own commands
A command is a prompt kept in a Markdown file, run as /name. Whatever you type after the name goes into the text. A subfolder makes folder:name.
- Yours go in the
commandsfolder of your data folder. - A project's go in
.letronna/commandsor.claude/commands.
The files are the same as Claude Code's: name.md, with an optional description and argument-hint in the frontmatter.
When you run a command, its text is put in the box for you to read before you send it. If a project's command has the same name as yours, yours wins.
Saved workflows run as slash commands too, and /review comes built in.