# Mindset Skills

> Write a policy or process once and share it with every agent that needs it. Agents read it only when a request calls for it.

Builders work from a coding agent and operators from an AI client, both through the MCP server, or in the Studio. In Mindset, an agent is a prompt and a model, held to an order by a script when needed. It uses connections to outside systems, operations on them, functions and widgets, and loads Skills when a request calls for them. Every change is a version, and every execution is a run. Every piece sits on one security and compliance layer: the org sets its posture once, and every agent works under it, with the audit trail and, once an admin turns it on, personal data redaction. The platform runs in your own cloud (AWS, GCP or Azure) or is hosted by Mindset. Your app reaches agents through the SDK, your server through the API, and people through the Hub and schedules. Highlighted: Skill.

A **Mindset Skill** is a set of instructions your [agents](https://docs.mindset.ai/docs/agents) load when a request calls for it: a returns policy, your brand voice, the way your team marks an assessment. Use one for context an agent needs that doesn't belong in its prompt, because it changes by customer or situation. You write the Skill once, attach it to the agents that need it, and change it in one place.

Skills use the open Agent Skills format, a `SKILL.md` with text files beside it, so a skill your team already wrote for any coding agent can come across, as long as it has no scripts.

## Why it matters

When a policy changes, every agent that holds its Skill follows the new version from its next turn, with no new agent version. An agent can also hold long reference material without carrying it on every turn, because it reads a Skill only when a request needs it. That reference can be how to call a particular API, details about a customer's organization, or anything else an agent needs only some of the time.

## How an agent uses one

On every turn, an agent sees a short list of the Skills it holds, each with its name and a description of when to use it. When a request matches a description, the agent loads that Skill's instructions and follows them as your org's own. If the instructions point to a file, such as a set of email templates, it reads that file too.

This makes the description the most important line in a Skill: along with the name, it's all the agent reads before deciding. [Add a Mindset Skill](https://docs.mindset.ai/docs/add-a-skill) covers how to write one that gets loaded.

Loading a Skill shows nothing in the conversation. Each load is recorded on the [run](https://docs.mindset.ai/docs/runs), and the Skill's **Overview** in the Studio counts its loads over the last 30 days.

> **Info:** The [skills on this site](https://docs.mindset.ai/skills) are a different thing. Your coding agent loads those to build with Mindset, while Mindset Skills are loaded by your agents as they run.

## When to use a Skill

| What you're adding                                                                                                                               | Where it goes                                                                                                                         |
| ------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------- |
| Know-how one or more agents need, such as how to word a supplier query, how to call a particular API, or details about a customer's organization | A Mindset Skill                                                                                                                       |
| Long reference an agent needs for only some requests                                                                                             | A Mindset Skill, with the detail in its files                                                                                         |
| One agent's role and behavior                                                                                                                    | That agent's prompt                                                                                                                   |
| A rule that must hold every time                                                                                                                 | A [script](https://docs.mindset.ai/docs/scripts) or a [function](https://docs.mindset.ai/docs/functions). An agent follows a Skill's instructions, but Mindset doesn't enforce them |

## The invoice exceptions agent

The finance team's rules for querying invoice differences live in a `supplier-queries` Skill: when a difference is worth raising with the supplier, and how to word the email, with the templates in `references/templates.md`. When the agent finds a difference, it loads the Skill and drafts the query the way finance would. When an invoice matches, it never opens it.

When finance changes a template, that's one new version of the Skill. Once it's activated, the agent uses it from its next turn, and so does any other agent that holds it.

To write and attach one, see [Add a Mindset Skill](https://docs.mindset.ai/docs/add-a-skill).
