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

# Skills & Skills Review

> Reusable how-to instructions for agents, and the admin Skills Review queue.

A **skill** is a reusable block of instructions — a how-to your agents (and you, in Chat) can invoke by name. Where knowledge says *what is true*, a skill says *how to do something*.

**Where:** Sidebar › Knowledge › **Skills**. The admin review queue is at Sidebar › Knowledge › **Skills Review**.

## The Skills view

A standard table view over your skill library (see [Data Views](/guide/data-views) for table controls), filtered by your active department and each skill's scope.

## Anatomy of a skill

| Field            | Meaning                                                                                  |
| ---------------- | ---------------------------------------------------------------------------------------- |
| **Name**         | How the skill is invoked and displayed                                                   |
| **Category**     | From your company's skill categories (color-coded in Chat)                               |
| **Scope**        | user / department / global — who can see and use it                                      |
| **Department**   | When department-scoped                                                                   |
| **Prompt block** | The skill's content — the instructions injected into the prompt when the skill is active |

## Creating skills

* **In Chat:** click **Save as Skill** under any assistant reply — a confirm card opens with the fields above.
* **In the view:** the **Add** button creates one directly.
* **By agents:** agents can propose skill updates; proposals go through approval.

## Using skills

* **In Chat:** activate skills via the context picker's Skills tab or an `@`-mention; active skills show as colored chips above the composer.
* **On agents:** attach skills under Agents › Skills & Tools › Skills — the agent then carries those instructions on every run. Some skills auto-inject based on the agent's configuration (e.g. the Telegram delivery skill).
* **In workflows:** a Workflow agent can include a **Skill** step.

## Skills Review *(Company Admins)*

**Where:** Sidebar › Knowledge › **Skills Review**. Members see this entry locked.

The review queue for **promotion proposals** — skills, knowledge, and memory entries that the learning loop (the nightly curator sweep) proposes making permanent. Each pending card shows the proposed entry, its target (skill / knowledge / memory), scope, who or what requested it, and when.

* **Approve** — applies the proposal to the target library.
* **Reject** — discards it and lowers the confidence of the source memory, so the same thing isn't proposed again.

The same proposals also surface in the admin's [Inbox](/guide/inbox); deciding in either place resolves them.