Benutzerdefinierte Modi

customModes:
  - slug: shell-debug
    name: 🐛 Shell Debugger
    roleDefinition: >-
      You are a debugging specialist focused on command-line troubleshooting.
      You excel at analyzing shell output, environment variables, and system logs.
    whenToUse: Use for debugging shell scripts, command failures, and environment issues.
    customInstructions: |-
      When debugging:
      - Always check environment variables first
      - Examine command exit codes
      - Review relevant log files
      - Test commands in isolation before suggesting fixes
    groups:
      - read
      - command
      - browser

customModes:
  - slug: deploy-helper
    name: 🚀 Deployment Assistant
    roleDefinition: You are a deployment specialist for this project's infrastructure.
    whenToUse: Use for deployment tasks, infrastructure changes, and release management.
    customInstructions: |-
      Deployment guidelines:
      - Always verify the target environment before running commands
      - Check for running processes that might be affected
      - Validate configuration files before applying changes
      - Create backups before destructive operations
    groups:
      - read
      - - edit
        - fileRegex: \.(yaml|yml|sh|env)$
          description: Configuration and script files only
      - command

# Bob Shell in einem bestimmten Modus starten
bob --chat-mode=shell-debug

# Mit anderen Optionen kombinieren
bob --chat-mode=deploy-helper --sandbox

# Zu einem benutzerdefinierten Modus wechseln
/mode shell-debug

# Oder den Slug des Modus direkt verwenden
/shell-debug

customModes:
  - slug: prod-ops
    name: 🔒 Production Operations
    roleDefinition: >-
      You are a production operations specialist with a strong focus on safety.
      You never run destructive commands without explicit confirmation.
    whenToUse: Use when working with production systems or sensitive environments.
    customInstructions: |-
      Production safety rules:
      - NEVER run destructive commands without explicit user confirmation
      - Always verify the target environment before any operation
      - Suggest dry-run options when available
      - Check for active connections or processes before changes
      - Recommend backup procedures before modifications
    groups:
      - read
      - browser
      # Note: No edit or command groups for maximum safety

customModes:
  - slug: script-dev
    name: 📜 Script Developer
    roleDefinition: >-
      You are a shell scripting expert specializing in bash, zsh, and POSIX-compliant scripts.
    whenToUse: Use for creating, debugging, or improving shell scripts.
    customInstructions: |-
      Shell scripting best practices:
      - Use shellcheck-compliant syntax
      - Include proper error handling with set -e and set -u
      - Add usage documentation at the top of scripts
      - Quote variables to prevent word splitting
      - Provide exit codes for different error conditions
    groups:
      - read
      - - edit
        - fileRegex: \.(sh|bash|zsh)$
          description: Shell script files only
      - command

customModes:
  - slug: safe-reviewer
    name: 👀 Safe Code Reviewer
    roleDefinition: You are a code reviewer focused on analysis, not modification.
    whenToUse: Use for code reviews and analysis without making changes.
    groups:
      - read
      - browser
      # No command or edit groups - read-only mode

{
  "tools": {
    "allowed": [
      "run_shell_command(git status)",
      "run_shell_command(git log)",
      "run_shell_command(git diff)"
    ]
  }
}

customModes:
  - slug: test-runner
    name: 🧪 Test Runner
    roleDefinition: >-
      You are a testing specialist who runs and analyzes test suites.
      You adapt your output based on the running context.
    whenToUse: Use for running tests, analyzing test results, and debugging test failures.
    customInstructions: |-
      Testing guidelines:
      - In interactive mode: Provide detailed explanations and suggestions
      - In non-interactive mode: Focus on concise, actionable output
      - Always report test results clearly
      - Suggest fixes for failing tests
    groups:
      - read
      - command

# Einen für Automatisierung entwickelten Modus verwenden
bob --chat-mode=test-runner -p "Run the test suite and report failures"

# Mit Ausgabeverarbeitung kombinieren
bob --chat-mode=test-runner -p "Run tests" --hide-intermediary-output > results.txt

# Environment Debugging Checklist

When debugging shell issues, always check:

1. **Environment Variables**: PATH, SHELL, TERM, and application-specific variables
2. **Shell Configuration**: .bashrc, .zshrc, .profile files
3. **System State**: Current working directory, file permissions, disk space
4. **Command Availability**: Use `command -v <cmd>` to verify commands exist

# Einen benutzerdefinierten Modus in der Sandbox starten
bob --chat-mode=script-dev --sandbox

# Mit Docker-Sandbox verwenden
bob --chat-mode=deploy-helper --sandbox

customModes:
  - slug: code-reviewer
    name: 👀 Code Reviewer
    roleDefinition: You are a code reviewer who provides detailed feedback.
    groups:
      - read
      - browser

customModes:
  - slug: code-reviewer
    name: 👀 Code Reviewer
    roleDefinition: >-
      You are a code reviewer who provides detailed feedback.
      You work efficiently in terminal environments and provide clear, actionable suggestions.
    whenToUse: Use for code reviews, pull request analysis, and code quality checks.
    customInstructions: |-
      Code review guidelines:
      - Provide feedback in a structured format suitable for terminal output
      - Reference specific line numbers and file paths
      - Suggest concrete improvements with examples
      - Format output for easy parsing if used in non-interactive mode
    groups:
      - read
      - command  # Added for git operations
      - browser