Skip to main content

​
About Custom Commands

Custom commands are user-defined slash commands stored as .md files that appear in the / autocomplete menu in chat. They let you create reusable prompts for common workflows — code reviews, deployment checklists, bug fix templates, and more.
Custom Commands are available as a Public Beta opt-in from the Settings page in VSCode 0.789.0+. They are also being rolled out server-side for Community (Vanguard) tier users.

​
Creating a Command

Create a markdown file in one of the supported command directories: 
# Workspace command (shared with your team)
mkdir -p .augment/commands
echo "Review this code for security vulnerabilities and suggest fixes." > .augment/commands/security-review.md

# User command (available across all projects)
mkdir -p ~/.augment/commands
echo "Analyze this code for performance issues:" > ~/.augment/commands/optimize.md

​
Command Locations

Commands are discovered from multiple locations in order of precedence:
LocationScopeDescription
~/.augment/commands/UserAvailable across all workspaces (highest precedence)
.augment/commands/WorkspaceProject-specific, can be version controlled
.claude/commands/WorkspaceCompatible with Claude Code
~/.claude/commands/UserCompatible with Claude Code
.cursor/commands/WorkspaceCompatible with Cursor
When commands with the same name exist in multiple locations, the command from the higher-precedence location is used.

​
Using Commands

Type / in the chat input to open the autocomplete menu, then select or search for your command. You can also type the command name directly: 
/security-review
Arguments can be passed after the command name: 
/fix-issue 123

​
Frontmatter

Command files support optional YAML frontmatter for metadata:
FieldDescriptionDefault
descriptionBrief description shown in the autocomplete menuFirst line of the prompt
argument-hintExpected arguments format displayed after the command nameNone
modelModel to use for this command (overrides default)Default model
Example — .augment/commands/deploy-staging.md: 
---
description: Deploy the application to staging with health checks
argument-hint: [branch-name]
model: gpt-4o
---

Deploy the application to the staging environment:

1. Run all tests to ensure code quality
2. Build the application for production
3. Deploy to staging server
4. Run health checks to verify deployment
5. Send notification to team channel

​
Template Substitution

Commands support template variables for dynamic content:
PatternDescription
$ARGUMENTSReplaced with any arguments passed after the command name
{placeholder}Prompts the user for a value
{{placeholder}}Alternative syntax for user prompts
Example using $ARGUMENTS: 
---
description: Fix a GitHub issue
argument-hint: [issue-number]
---

Fix the GitHub issue described below. Follow our coding standards and include tests.

Issue: $ARGUMENTS

​
Namespacing

Organize commands in subdirectories using the namespace:command syntax: 
.augment/commands/
  ├── deploy-staging.md          → /deploy-staging
  ├── frontend/
  │   ├── component.md           → /frontend:component
  │   └── test.md                → /frontend:test
  └── backend/
      └── migrate.md             → /backend:migrate

​
Best Practices

  1. Use kebab-case naming — e.g., deploy-staging.md, run-tests.md
  2. Include clear descriptions in frontmatter for better discoverability
  3. Use workspace commands for team-shared workflows
  4. Use user commands for personal preferences across projects
  5. Organize with subdirectories for related command groups
  6. Break complex workflows into numbered steps

​
See Also