[Docs index](/docs.md) / Agent Skills

---

# Agent Skills

Agent Skills help mid-sized companies clean up scattered AI usage by moving reusable prompts, team instructions, and informal AI workflows into Assist as reviewed, versioned procedures.

Use Agent Skills when employees are already using AI in many places: private chat accounts, copied prompt docs, Slack threads, spreadsheets, department wikis, and one-off experiments. Assist gives those workflows a shared home where teams can review the instructions, attach them to operational agents, and keep track of which version is live.

## Why teams use Agent Skills

Mid-sized companies usually adopt AI before they have a central operating model for it. A support lead may have a strong escalation prompt in a doc. A sales manager may have a discovery-call summary prompt. Operations may have a weekly reporting prompt. Finance may have a careful invoice-review checklist. These workflows can be useful, but they are hard to govern when they live in personal accounts and disconnected notes.

Agent Skills turn those scattered practices into reusable Assist instructions. Each skill has a main `/SKILL.md` file, optional supporting files, a draft and approval lifecycle, and a usage view that shows where the approved version is attached.

## Good first use cases

- Inventorying team AI usage and deciding what should move into Assist.
- Standardizing copied prompts that multiple employees already use.
- Creating reviewed instructions for support, sales, operations, finance, or customer success agents.
- Replacing informal prompt docs with versioned instructions and examples.
- Rolling out a department workflow without silently changing every agent at once.

## Who can create Agent Skills

- Workspace members can create workspace-scoped Agent Skills in their current workspace.
- Workspace members can create personal Agent Skills for private drafting and experiments.
- Workspace admins can review, approve, reject, and govern workspace Agent Skills.
- Workspace admins can see personal Agent Skills in their workspace for support and governance.

## Who can use Agent Skills

- An agent can use an Agent Skill after an approved version is attached to that agent.
- The person attaching the skill must be able to edit the target agent and read the skill.
- Agents stay pinned to the attached version until someone upgrades the attachment.
- An agent can read mounted skill files when the skill is relevant to the user's request.

## What agents can do with attached skills

- See a compact index of attached Agent Skills.
- Read the relevant `SKILL.md` procedure before following it.
- Read supporting checklists, prompt examples, templates, and policy notes.
- Search attached skill files with filesystem search tools.
- Keep using a pinned, reviewed version until the team upgrades it.

## What agents cannot do with attached skills

- Change a skill's files through the mounted path.
- Write new files into a skill attachment.
- Automatically move to a newer approved version.
- Use a skill they cannot read.
- Use a skill attached to another agent.
- Treat a draft or pending version as production-ready.

## How Agent Skills work

An Agent Skill has a name, slug, scope, and versions. Each version contains files. The main file is `/SKILL.md`, which tells the agent when to use the skill and how to handle the work. Supporting files can hold prompt examples, checklists, department rules, templates, redlines from reviewers, or sample outputs.

For a company trying to clean up AI usage, this matters because the goal is not only better prompts. The goal is a controlled path from "someone found a useful AI workflow" to "the team has a reviewed Assist procedure that can be reused, improved, and audited."

## Scopes

| Scope | Best for | Who can see it |
|-------|----------|----------------|
| Workspace | Shared AI workflows for support, sales, operations, finance, customer success, and leadership teams | Workspace members can read approved workspace skills. Editors and admins can manage versions. |
| Personal | Drafting, experiments, and private cleanup work before proposing a workflow to the team | The owner and workspace admins can see it. Other members cannot read it unless policy changes. |

## Lifecycle

1. Create a managed AI workflow skill or import an existing team workflow.
2. Author the draft files, especially `/SKILL.md`.
3. Submit the draft for review.
4. An admin approves or rejects the pending version.
5. Attach an approved version to one or more operational agents.
6. Upgrade attachments when a newer approved version should roll out.

## Supported import sources

Agent Skills can be imported from:

- A Skills Marketplace page that points to a GitHub repository.
- A GitHub repository URL.
- A GitHub folder URL.
- An `owner/repo` shorthand.
- A `skills add owner/repo` command.

Imported skills are installed as drafts. Review every file before submitting or attaching the skill to an agent.

## Related guides

- [Use cases and playbooks](use-cases.md)
- [Creating a managed AI workflow skill](creating-a-managed-ai-workflow-skill.md)
- [Importing a team AI workflow skill](importing-a-team-ai-workflow-skill.md)
- [Authoring and approving Agent Skills](authoring-and-approving-agent-skills.md)
- [Attaching skills to operational agents](attaching-skills-to-operational-agents.md)
- [Governing Agent Skill versions](governing-agent-skill-versions.md)
- [Troubleshooting](troubleshooting.md)

---

## Navigation

### In this section: Agent Skills

- **Agent Skills** (current)
- [Use Cases and Playbooks](/docs/agent-skills/use-cases.md)
- [Troubleshooting Agent Skills](/docs/agent-skills/troubleshooting.md)
- [Attaching Agent Skills to Operational Agents](/docs/agent-skills/attaching-skills-to-operational-agents.md)
- [Authoring and Approving Agent Skills](/docs/agent-skills/authoring-and-approving-agent-skills.md)
- [Creating a Managed AI Workflow Skill](/docs/agent-skills/creating-a-managed-ai-workflow-skill.md)
- [Governing Agent Skill Versions](/docs/agent-skills/governing-agent-skill-versions.md)
- [Importing a Team AI Workflow Skill](/docs/agent-skills/importing-a-team-ai-workflow-skill.md)

#### Playbooks

- [Playbook: Build a Department AI Rollout Skill](/docs/agent-skills/playbook-department-ai-rollout.md)
- [Playbook: Build a Team AI Usage Cleanup Skill](/docs/agent-skills/playbook-ai-usage-inventory.md)
- [Playbook: Build a Team Prompt Standardization Skill](/docs/agent-skills/playbook-prompt-standardization.md)

### Other sections

- [MCP Servers](/docs/mcp-servers/overview.md)
- [Tool Creation](/docs/tool-creation/overview.md)
- [Agent Filesystem](/docs/agent-filesystem/overview.md)
- [Chat Sharing](/docs/chat-sharing/overview.md)
- [Scheduled Triggers](/docs/scheduled-triggers/overview.md)
- [Sandcastles](/docs/sandcastles/overview.md)
- [Subagents](/docs/subagents/overview.md)
- [Workspace Permissions](/docs/workspace-permissions/overview.md)

[Back to docs index](/docs.md)