Skills and Slash Commands

Skills are reusable, structured instructions and domain knowledge that agents can draw on. They are named, and each skill's name doubles as a slash command you can type in the AI Chat input to invoke it directly.

This article reflects the current SKILL.html-based format, shipped via the Studio AI Core Artifacts extension. It replaces an earlier .prompttemplate-based format described in older versions of this article; if you have old-format skill files, they will not appear in the current Skills list.

What Is a Skill?

A skill is a self-contained folder with a SKILL.html entry point, plus any supporting reference files (and, in some cases, evaluation examples). SKILL.html starts with a YAML frontmatter block and a markdown body:

---
name: my-skill
description: One or two sentences describing what this skill does and when to use it.
---

Detailed instructions for the task, written in markdown.

The skill's name is also its slash command trigger; the skill above would be invoked with /my-skill. Studio AI can also load a relevant skill automatically based on your request, without you typing the slash command yourself, when the addressed agent (such as @k2-assistant) is configured to do so.

Using Slash Commands in the Chat

Type / in the chat input to see a list of available skills as slash commands. Select one, or type the full name and press Enter.

/fabric-commands

Skills Tab (AI Configuration)

Every installed skill is listed in the Skills tab of AI Configuration: its name, description, and the file path to its SKILL.html. Click Open next to a skill to view its file directly.

Where Skills Are Stored

Skills live under .agents/skills/<skill-name>/SKILL.html in your project (imported by the Studio AI Core Artifacts extension; see Getting Started with Studio AI). The @CreateSkill agent also supports an alternate .prompts/skills/ location. Check the Location column in the Skills tab for the exact path of any given skill rather than assuming one location.

Skills Available in This Project

At the time of writing, this project's Skills tab listed the following (yours may differ):

Skill Slash Command
broadway-actor-builder/broadway-actor-builder
broadway-flow-builder/broadway-flow-builder
data-product-builder/data-product-builder
data-product-cowork/data-product-cowork
debug-broadway-flow/debug-broadway-flow
fabric-commands/fabric-commands
fabric-java-docs/fabric-java-docs
fabric-overview/fabric-overview
fabric-project-helper/fabric-project-helper
fabric-troubleshooting/fabric-troubleshooting
interface-builder/interface-builder
report-builder/report-builder
web-service-builder/web-service-builder

Earlier built-in commands such as /remember, /with-apptester, /analyze-gh-ticket, /fix-gh-ticket, and /address-gh-review were not found in the current slash-command list and appear to have been replaced by this skill-based system.

Creating Custom Skills

To create a skill manually, add a folder under .agents/skills/<skill-name>/ containing a SKILL.html file with the frontmatter and body format shown above. Reload the Studio window (F1 > Developer: Reload Window) after adding a new skill file to make it available.

Alternatively, address @CreateSkill and describe the skill you want; it will scaffold a well-structured SKILL.html for you. See Studio AI Agents and Other Studio AI Agents.

Example: A Code Review Skill

---
name: k2review
description: Review the current file for K2View best practices and common issues, such as JDBC resource management, SQL injection risks, exception handling, and naming conventions.
---

Review the code in {{currentRelativeFilePath}} for the following issues:

1. JDBC resource management: ensure all connections, statements, and result sets are
   closed, preferably with try-with-resources.
2. SQL injection risks: verify that all SQL uses parameterized queries.
3. Exception handling: check that exceptions are caught at the appropriate level and
   logged correctly per our conventions.
4. Naming conventions: verify method and variable names follow our team standards.

For each issue found, explain the problem and provide a corrected version of the relevant code.
If no issues are found in a category, say so explicitly.

With this skill saved under .agents/skills/k2review/SKILL.html, typing /k2review in the chat triggers the full review prompt against the currently open file.

Skills and Slash Commands

Skills are reusable, structured instructions and domain knowledge that agents can draw on. They are named, and each skill's name doubles as a slash command you can type in the AI Chat input to invoke it directly.

This article reflects the current SKILL.html-based format, shipped via the Studio AI Core Artifacts extension. It replaces an earlier .prompttemplate-based format described in older versions of this article; if you have old-format skill files, they will not appear in the current Skills list.

What Is a Skill?

A skill is a self-contained folder with a SKILL.html entry point, plus any supporting reference files (and, in some cases, evaluation examples). SKILL.html starts with a YAML frontmatter block and a markdown body:

---
name: my-skill
description: One or two sentences describing what this skill does and when to use it.
---

Detailed instructions for the task, written in markdown.

The skill's name is also its slash command trigger; the skill above would be invoked with /my-skill. Studio AI can also load a relevant skill automatically based on your request, without you typing the slash command yourself, when the addressed agent (such as @k2-assistant) is configured to do so.

Using Slash Commands in the Chat

Type / in the chat input to see a list of available skills as slash commands. Select one, or type the full name and press Enter.

/fabric-commands

Skills Tab (AI Configuration)

Every installed skill is listed in the Skills tab of AI Configuration: its name, description, and the file path to its SKILL.html. Click Open next to a skill to view its file directly.

Where Skills Are Stored

Skills live under .agents/skills/<skill-name>/SKILL.html in your project (imported by the Studio AI Core Artifacts extension; see Getting Started with Studio AI). The @CreateSkill agent also supports an alternate .prompts/skills/ location. Check the Location column in the Skills tab for the exact path of any given skill rather than assuming one location.

Skills Available in This Project

At the time of writing, this project's Skills tab listed the following (yours may differ):

Skill Slash Command
broadway-actor-builder/broadway-actor-builder
broadway-flow-builder/broadway-flow-builder
data-product-builder/data-product-builder
data-product-cowork/data-product-cowork
debug-broadway-flow/debug-broadway-flow
fabric-commands/fabric-commands
fabric-java-docs/fabric-java-docs
fabric-overview/fabric-overview
fabric-project-helper/fabric-project-helper
fabric-troubleshooting/fabric-troubleshooting
interface-builder/interface-builder
report-builder/report-builder
web-service-builder/web-service-builder

Earlier built-in commands such as /remember, /with-apptester, /analyze-gh-ticket, /fix-gh-ticket, and /address-gh-review were not found in the current slash-command list and appear to have been replaced by this skill-based system.

Creating Custom Skills

To create a skill manually, add a folder under .agents/skills/<skill-name>/ containing a SKILL.html file with the frontmatter and body format shown above. Reload the Studio window (F1 > Developer: Reload Window) after adding a new skill file to make it available.

Alternatively, address @CreateSkill and describe the skill you want; it will scaffold a well-structured SKILL.html for you. See Studio AI Agents and Other Studio AI Agents.

Example: A Code Review Skill

---
name: k2review
description: Review the current file for K2View best practices and common issues, such as JDBC resource management, SQL injection risks, exception handling, and naming conventions.
---

Review the code in {{currentRelativeFilePath}} for the following issues:

1. JDBC resource management: ensure all connections, statements, and result sets are
   closed, preferably with try-with-resources.
2. SQL injection risks: verify that all SQL uses parameterized queries.
3. Exception handling: check that exceptions are caught at the appropriate level and
   logged correctly per our conventions.
4. Naming conventions: verify method and variable names follow our team standards.

For each issue found, explain the problem and provide a corrected version of the relevant code.
If no issues are found in a category, say so explicitly.

With this skill saved under .agents/skills/k2review/SKILL.html, typing /k2review in the chat triggers the full review prompt against the currently open file.