> For the complete documentation index, see [llms.txt](https://docs.darcyiq.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.darcyiq.com/core-features/skills.md).

# Skills

Teach DarcyIQ reusable expertise that persists across conversations. Skills are instruction sets — written in markdown — that tell the AI **how** to perform a specific task, follow a process, or apply domain knowledge. Create them yourself, ask Darcy to build one in chat, or import them using the open [agentskills.io](https://agentskills.io) standard. **Skills also lets you attach reference files — templates, scripts, datasets — that Darcy can read whenever the skill runs, and you can publish your best skills to a shared** [**Skills Catalog**](#skills-catalog) **so your whole organization can install them with one click.**

{% hint style="success" %}
**Example:** Save a "Client Proposal Writer" skill with your company's formatting rules, pricing structure, and tone guidelines, plus a proposal template and a pricing sheet attached as files. Every future conversation can load that skill instantly — no re-explaining required.
{% endhint %}

## Why Skills?

Skills solve a common problem with AI assistants: **repeating the same instructions across conversations**. Instead of re-typing how you want proposals formatted, how your deployment pipeline works, or what your company's naming conventions are, you save those instructions once as a skill and DarcyIQ loads them on demand.

| Feature              | Description                                                                                     |
| -------------------- | ----------------------------------------------------------------------------------------------- |
| **Persistent**       | Saved to your account and available across all conversations                                    |
| **On-Demand**        | Loaded only when needed, keeping conversations fast and focused                                 |
| **File Attachments** | Attach templates, scripts, and reference docs that Darcy reads alongside the skill instructions |
| **Cross-Platform**   | Import SKILL.md packages from Anthropic, ChatGPT, and other agent platforms                     |
| **Versioned**        | Every edit increments the version so you can track changes and roll back                        |
| **Shareable**        | Publish to your organization's Skills Catalog for one-click install by any teammate             |
| **Toggleable**       | Enable or disable skills without deleting them                                                  |
| **Import / Export**  | Share skills using standard SKILL.md files compatible with agentskills.io                       |

## Creating Skills

### From the Skills Panel

{% stepper %}
{% step %}
**Open the Skills Panel** Click the **Skills** icon (lightbulb) in the sidebar to open the Skills drawer.
{% endstep %}

{% step %}
**Click "New"** Click the **New** button in the top-right corner of the panel.
{% endstep %}

{% step %}
**Fill in the Details**

| Field           | Description                                                   | Limits             |
| --------------- | ------------------------------------------------------------- | ------------------ |
| **Name**        | A descriptive name (auto-converted to a URL-safe slug)        | 2–64 chars         |
| **Description** | Brief summary of what the skill does                          | Up to 500 chars    |
| **Skill Body**  | The full instructions in markdown — this is what the AI reads | Up to 15,000 chars |
| {% endstep %}   |                                                               |                    |

{% step %}
**Save** Click **Create** to save. The skill is immediately available in all conversations.
{% endstep %}
{% endstepper %}

### In Chat

Ask DarcyIQ to create a skill directly in conversation:

* "Create a skill called `meeting-summary` that formats meeting notes with attendees, action items, and next steps."
* "Save what we just discussed as a skill I can reuse."
* "Make a skill for writing customer onboarding emails in our company's tone."

DarcyIQ saves the skill to your account automatically. You can view and manage it from the Skills panel afterwards.

### Import a SKILL.md File

If you have a skill file from another user, from the [agentskills.io](https://agentskills.io) community, or from another agent platform:

1. Open the **Skills Panel**
2. Click **Import**
3. Select a `.md` file in the SKILL.md format
4. Review the imported name, description, and body
5. Click **Create** to save

{% hint style="info" %}
**Cross-platform compatibility**: DarcyIQ follows the open [agentskills.io](https://agentskills.io) standard, so SKILL.md packages built for Anthropic Claude, ChatGPT, and other agent platforms can be imported and used directly — no rework required. Any file that uses YAML frontmatter with `name` and `description` fields followed by a markdown body is compatible.
{% endhint %}

## File Attachments

Skills can include reference files — templates, scripts, datasets, schemas, examples, or anything else Darcy might need to do the task well. When the skill is active, those files are available to Darcy in the conversation's workspace.

| Capability              | Detail                                                                                                                                     |
| ----------------------- | ------------------------------------------------------------------------------------------------------------------------------------------ |
| **File types**          | Markdown, text, PDFs, Office docs, images, code files, JSON/YAML/CSV — most formats                                                        |
| **Per-file size limit** | 5 MB                                                                                                                                       |
| **Files per skill**     | Up to 100                                                                                                                                  |
| **Storage**             | Files are stored alongside the skill and travel with it                                                                                    |
| **Workspace staging**   | Available to Darcy through [DarcyOS](/core-features/darcy-os.md) — Darcy can read them, run scripts you attached, or use them as templates |

### Adding Files to a Skill

{% stepper %}
{% step %}
**Open the Skill** From the Skills panel, click any skill to open the editor.
{% endstep %}

{% step %}
**Expand the Files Section** Scroll to the **Files** section in the editor and click to expand it.
{% endstep %}

{% step %}
**Upload Files** Click **Upload** and select one or more files from your machine. Files appear in the list as soon as they finish uploading.
{% endstep %}

{% step %}
**Save** The skill (and its files) are saved to your account and ready to use in any conversation.
{% endstep %}
{% endstepper %}

### Managing Attached Files

| Action       | How                                                             |
| ------------ | --------------------------------------------------------------- |
| **List**     | Open the skill editor and expand the Files section              |
| **Download** | Click the download icon next to any file                        |
| **Delete**   | Click the trash icon (the file is removed from the skill only)  |
| **Replace**  | Upload a file with the same name to overwrite the existing copy |

### Using Attached Files

When a skill with attachments is active in a conversation, Darcy can read those files automatically as part of executing the skill. Common patterns:

* **Templates** — proposal, SOW, or report templates that Darcy fills in
* **Reference data** — pricing sheets, product catalogs, lookup tables
* **Scripts** — helper scripts Darcy can run on [DarcyOS](/core-features/darcy-os.md)
* **Examples** — sample outputs that show Darcy what "good" looks like
* **Schemas** — JSON/YAML schemas Darcy validates against

{% hint style="info" %}
**Pro Tip**: For binary files (DOCX, PDF, images, scripts), Darcy uses [DarcyOS](/core-features/darcy-os.md) to read them directly from the workspace. This means a skill with a Python helper script can have Darcy actually run that script when the skill is invoked.
{% endhint %}

## Managing Skills

### Skills Panel

| Action             | How                                                 |
| ------------------ | --------------------------------------------------- |
| **View all**       | Open the Skills panel from the sidebar              |
| **Edit**           | Click any skill to open the editor                  |
| **Enable/Disable** | Toggle the switch next to any skill                 |
| **Export**         | Click the download icon to get a SKILL.md file      |
| **Delete**         | Click the trash icon (permanent — cannot be undone) |
| **Import**         | Click Import and select a SKILL.md file             |

### In Chat

You can also ask DarcyIQ to manage skills conversationally:

* "Update my `deploy-checklist` skill with these new steps..."
* "In my `deploy-checklist` skill, change 'staging' to 'pre-prod'"
* "Delete the `old-template` skill"
* "What skills do I have?"

{% hint style="warning" %}
**Disabling vs. Deleting**: Use the toggle switch to temporarily disable a skill without losing it. Deletion is permanent and cannot be undone.
{% endhint %}

## Skills Catalog

The Skills Catalog turns individual expertise into shared organizational capability. Instead of emailing SKILL.md files or recreating workflows from scratch, publish a skill once and let every teammate discover and install it with a single click — then keep everyone in sync when you publish updates.

The catalog lives right inside the Skills panel. Switch between the **My Skills** and **Catalog** tabs at the top of the panel to move between your own skills and everything published across your organization.

{% hint style="success" %}
**Example:** Your solutions architect builds an "AWS Migration Assessment" skill and publishes it to the catalog. Every consultant in the org installs it in one click — and when the architect refines the methodology, everyone is prompted to update to the latest version.
{% endhint %}

### Publishing a Skill

Publishing takes the **latest saved version** of a skill — its instructions *and* attached files — and creates a stable, shareable snapshot in the catalog. Installers always get that published snapshot, so your work-in-progress edits stay private until you publish them.

{% stepper %}
{% step %}
**Open the skill's actions** In the **My Skills** tab, hover a skill you own and open its actions menu (or open the skill editor).
{% endstep %}

{% step %}
**Click "Publish to Catalog"** The current saved version is snapshotted and listed in your organization's catalog. Your name is shown as the author.
{% endstep %}

{% step %}
**Keep it current** After making further edits, choose **Update Catalog Version** to push your latest saved version to the catalog. To pull a listing down entirely, choose **Unpublish**.
{% endstep %}
{% endstepper %}

{% hint style="info" %}
**Who can publish**: Publishing, updating a catalog listing, and unpublishing all require skill ownership **plus** the `Publish to Skill Library` permission. This is an opt-in permission that administrators grant through role management, so your organization controls exactly who can share skills org-wide.
{% endhint %}

{% hint style="warning" %}
**Publish after uploading files**: A published snapshot only includes files that existed when the version was saved. If you add attachments to a skill, save a new version and then publish (or update the catalog version) so installers receive those files.
{% endhint %}

### Installing a Skill

Anyone with skill access can browse the **Catalog** tab, search by name or description, and install what they need.

* **One-click install** — Click a catalog skill to add it to your own **My Skills** list. Darcy copies the instructions and all attached files into a brand-new skill that belongs to you — edit it freely without affecting the original.
* **No duplicates** — You can't install the same catalog skill twice, and you can't install your own published skill. If a name collides with an existing skill, Darcy automatically gives your copy a unique name.
* **Update notifications** — When the author publishes a newer version, your installed copy shows that an update is available. Click **Update** to overwrite your copy with the latest published snapshot in one step.

### Version Control & Rollbacks

Every catalog action is captured in a skill's version history, so nothing is ever lost:

* Installing a catalog skill creates an initial **`catalog_install`** version snapshot.
* Updating an installed skill from the catalog creates a **`catalog_update`** snapshot before applying the new content.

Because each of these is a full snapshot, you can open the skill's version history and **roll back** to a previous version at any time if an update doesn't work as expected. If you archive an installed catalog skill, it also appears under the archived view with a **Restore** action.

## Using Skills in Conversations

DarcyIQ is automatically aware of your active skills at the start of every conversation. You don't need to list or reference them — DarcyIQ already knows what's available.

To use a skill, simply ask:

* "Use my `client-proposal-writer` skill to draft a proposal for Acme Corp."
* "Follow my `meeting-summary` skill to format these notes."

DarcyIQ may also load a relevant skill on its own when it recognizes your request matches one.

### Chat Quick-Toggle

The chat input area includes a skills toggle where you can enable or disable specific skills before sending a message — no need to open the full Skills panel.

## Integration with Agents

Skills work alongside [Agents](/core-features/agents.md). When you chat with an agent that has its own custom instructions, your skills layer additional expertise on top of the agent's configuration. Skills are also available during scheduled agent runs.

## Use Cases

| Category                | Examples                                                               |
| ----------------------- | ---------------------------------------------------------------------- |
| **Process & Templates** | Proposal formatting, meeting note structure, code review checklists    |
| **Domain Knowledge**    | Product specs, compliance rules, client account context                |
| **Workflow Automation** | Deployment checklists, incident response procedures, onboarding steps  |
| **Communication**       | Email templates, tone guidelines, client-specific language preferences |

## Best Practices

1. **Be specific**: "Format prices as USD with two decimal places" beats "Format prices correctly"
2. **Use structure**: Break instructions into sections with headers and lists
3. **Include examples**: Show what good output looks like
4. **Keep it focused**: One skill per task or domain — don't combine unrelated instructions
5. **Start small**: Begin with one or two high-value skills for the processes you explain most often, then expand from there

{% hint style="info" %}
**Pro Tip**: If you find yourself repeating the same instructions across multiple conversations, that's a great candidate for a skill.
{% endhint %}


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.darcyiq.com/core-features/skills.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
