Commands

Commands are slash commands that users invoke in their AI assistant.

Canonical Format

type Command struct {
    Name         string   `yaml:"name"`
    Description  string   `yaml:"description"`
    Prompt       string   `yaml:"prompt"`
    AllowedTools []string `yaml:"allowed_tools,omitempty"`
    Model        string   `yaml:"model,omitempty"`
}

Creating a Command

Using Go

package main

import (
    "github.com/agentplexus/assistantkit/commands/core"
)

func main() {
    cmd := core.Command{
        Name:        "build",
        Description: "Build the project",
        Prompt: `Build the project using the appropriate build system.
Detect the project type and run the correct build command:
- Go: go build ./...
- Node.js: npm run build
- Python: python setup.py build`,
        AllowedTools: []string{"Bash", "Read"},
    }
}

Using Markdown

Create commands/build.md:

---
name: build
description: Build the project
allowed_tools:
  - Bash
  - Read
---

Build the project using the appropriate build system.
Detect the project type and run the correct build command:

- Go: `go build ./...`
- Node.js: `npm run build`
- Python: `python setup.py build`

Command Options

Field	Description	Required
name	Command name (used as /name)	Yes
description	Short description shown in help	Yes
prompt	Instructions for the AI	Yes
allowed_tools	Tools the command can use	No
model	Preferred model (sonnet, opus, haiku)	No

Tool Restrictions

You can limit which tools a command can use:

allowed_tools:
  - Read      # Read files
  - Write     # Write files
  - Edit      # Edit files
  - Bash      # Run shell commands
  - Glob      # Find files by pattern
  - Grep      # Search file contents

Security

Commands with Bash access can execute arbitrary shell commands. Consider limiting tools for commands that process untrusted input.

Assistant-Specific Output

Claude Code

---
name: build
description: Build the project
allowed_tools:
  - Bash
  - Read
---

Build the project using the appropriate build system...

Gemini CLI

name: build
description: Build the project
prompt: |
  Build the project using the appropriate build system...

OpenAI Codex

---
name: build
description: Build the project
---

Build the project using the appropriate build system...

Examples

Test Command

---
name: test
description: Run project tests
allowed_tools:
  - Bash
  - Read
---

Run the project's test suite:

1. Detect the project type
2. Run appropriate test command
3. Report results clearly

For Go: `go test -v ./...`
For Node.js: `npm test`
For Python: `pytest -v`

Deploy Command

---
name: deploy
description: Deploy to production
allowed_tools:
  - Bash
  - Read
model: opus
---

Deploy the application to production:

1. Verify all tests pass
2. Check for uncommitted changes
3. Run the deployment script
4. Verify deployment succeeded

IMPORTANT: Always confirm with user before deploying.