Skip to main content
This is the new Settings UI currently in Beta. It uses a form-based interface for agent configuration. For JSON editing, use the classic settings.
Agent profiles are used throughout Vibe Kanban wherever agents run: onboarding, default settings, attempt creation, and follow-ups. The default configuration appears pre-selected in the agent dropdown in the chat input.

​
Common Use Cases

Use CaseExample Configuration
Fast iterationDisable planning mode, use faster model
Complex tasksEnable planning mode, use advanced model
Autonomous workSkip permission prompts (use with caution)
Code reviewEnable approvals for all changes
Multi-instanceEnable Claude Code Router for parallel work

​
Key Concepts

Understanding these concepts helps you configure agents effectively:
When enabled, the agent first creates a detailed plan before writing code. This adds an extra step where you review and approve the approach before implementation begins. Useful for complex tasks where you want to validate the strategy, but adds overhead for simple changes.
Agents request permission before performing potentially destructive actions like deleting files, running shell commands, or modifying system configurations. Skipping these prompts (dangerously_skip_permissions) allows fully autonomous operation but removes safety guardrails.
Distributes requests across multiple Claude Code instances running in parallel. Useful when working on several tasks simultaneously or when you want to reduce wait times by load-balancing across instances.
Controls what the agent can access:
  • read-only - Can read files but not modify anything
  • workspace-write - Can modify files within the project directory
  • danger-full-access - Unrestricted file system access
Determines when the agent pauses for your confirmation:
  • untrusted - Asks before every action
  • on-failure - Only asks when something goes wrong
  • on-request - Only asks when explicitly requested
  • never - Fully autonomous (no confirmations)
Controls how much computational effort the model spends “thinking” before responding. Higher reasoning produces more thorough analysis but takes longer and uses more tokens. Lower reasoning is faster but may miss nuances in complex problems.

​
Configuration Access

Access agent profiles via Settings → Agents in the Workspaces UI.
Agent configuration form editor interface with finder-style layout
The interface uses a finder-style two-column layout:
  • Left column - Lists all available agents
  • Right column - Lists configurations for the selected agent

​
Managing Configurations

​
Creating a New Configuration

  1. Click the + button in the Configurations column
Configuration column showing the + button with Create new tooltip
  1. Enter a Configuration Name (e.g., PRODUCTION, DEVELOPMENT)
  2. Optionally select a configuration to Clone from, or start blank
  3. Click Create Configuration
Create New Configuration dialog with name field and clone option

​
Editing a Configuration

  1. Select an agent from the left column
  2. Select a configuration from the right column
  3. Edit the settings in the form below
  4. Click Save in the bottom bar

​
Deleting a Configuration

  1. Select the configuration to delete
  2. Click the delete button
  3. Confirm deletion
You cannot delete the only configuration for an agent. Each agent must have at least one configuration.

​
Setting a Default Configuration

Click the Default badge next to a configuration to make it the default for that agent. The default configuration:
  • Appears pre-selected in the agent dropdown when creating new attempts
  • Is used for follow-up messages in workspaces
  • Shows in the chat input toolbar as the current agent/variant

​
Agent Configuration Options

​
plan
boolean
Enable planning mode for complex tasks
​
claude_code_router
boolean
Route requests across multiple Claude Code instances
​
dangerously_skip_permissions
boolean
Skip permission prompts (use with caution)
View full CLI reference →

​
Universal Options

These options work across multiple agent types:
​
append_prompt
string | null
Text appended to the system prompt
Options prefixed with “dangerously_” bypass safety confirmations and can perform destructive actions. Use with extreme caution.

​
Using Agent Configurations

Once configured, your agent variants appear in the chat input toolbar, allowing quick switching between configurations.
Chat input showing agent variant dropdown with Default, Approvals, Opus, and Plan options

Default Configuration

Set your default agent and variant in Settings → General → Default Coding Agent for consistent behaviour across all attempts.

Per-Attempt Selection

Override defaults when creating attempts by selecting different agent/variant combinations in the attempt dialogue.
MCP (Model Context Protocol) servers are configured separately under Settings → MCP Servers but work alongside agent profiles to extend functionality.

Connecting MCP Servers

Configure MCP servers within Vibe Kanban for your coding agents

Classic Agent Configuration

Use the classic settings for JSON-based agent configuration