import { z } from 'zod';
// Schema for prompt arguments
const ExecuteLudusCmdArgsSchema = z.object({
command_intent: z.string().describe('What you want to accomplish with the CLI command'),
target_user: z.string().optional().describe('Target user for admin operations (leave empty for current user)'),
confirm_destructive: z.boolean().optional().default(false).describe('Confirmation for destructive operations')
});
export type ExecuteLudusCmdArgs = z.infer<typeof ExecuteLudusCmdArgsSchema>;
export async function handleExecuteLudusCmdPrompt(args: ExecuteLudusCmdArgs) {
const { command_intent, target_user, confirm_destructive = false } = args;
// Convert string input to proper boolean
const isDestructiveConfirmed = typeof confirm_destructive === 'string'
? ['true', 'TRUE', 't', 'T', '1', 'yes', 'YES', 'y', 'Y'].includes(confirm_destructive)
: confirm_destructive === true;
return {
messages: [
{
role: "user",
content: {
type: "text",
text: `Execute Ludus CLI command safely to: ${command_intent}
${target_user ? `\nTarget user: ${target_user}` : '\nTarget: Current user'}
${isDestructiveConfirmed ? '\nDESTRUCTIVE ACTION CONFIRMED by user' : ''}
LUDUS CLI EXECUTION SAFETY PROTOCOL:
Follow this protocol exactly to ensure safe and effective CLI command execution using the ludus_cli_execute tool.
STEP 1: SAFETY ASSESSMENT
- Analyze the intent: "${command_intent}"
- Determine if this is a:
* INFORMATIONAL command (help, status, list, info) → LOW RISK
* CONFIGURATION command (set, config, update) → MEDIUM RISK
* DEPLOYMENT command (deploy, start, build) → HIGH RISK
* DESTRUCTIVE command (destroy, delete, remove, abort) → CRITICAL RISK
STEP 2: TOOL PREFERENCE CHECK
Before using raw CLI, check if a dedicated MCP tool exists:
- Range deployment → use \`deploy_range\` tool instead
- Range destruction → use \`destroy_range\` tool instead
- Range status → use \`get_range_status\` tool instead
- User management → use \`list_user_ranges\` tool instead
- Configuration → use \`set_range_config\` tool instead
- Help information → use \`ludus_help\` tool instead
ONLY use \`ludus_cli_execute\` when:
- No dedicated tool exists for your specific need
- You need advanced CLI features not covered by tools
- Troubleshooting requires raw CLI access
- User explicitly requests direct CLI access
STEP 3: PRE-EXECUTION STATE CHECK
For all non-informational commands:
1. Use \`get_range_status\` to check current range state
2. Use \`list_user_ranges\` to understand existing resources
3. Document what will be affected by your command
STEP 4: DESTRUCTIVE ACTION PROTOCOL
For destructive commands (destroy, delete, remove, abort):
- MANDATORY: Explain exactly what will be destroyed/affected
- MANDATORY: Ask for explicit user confirmation
- MANDATORY: Wait for user to confirm before proceeding
- NEVER execute destructive commands without confirmation
- Set \`confirm_destructive: true\` only AFTER user confirms
Examples of DESTRUCTIVE commands requiring confirmation:
- \`range rm\` or \`range destroy\` → Deletes entire range permanently
- \`user rm\` → Removes user account and all data
- \`range abort\` → Stops deployment, may leave partial state
STEP 5: ADMIN OPERATION AWARENESS
${target_user ? `
ADMIN OPERATION DETECTED:
- Target user: "${target_user}"
- This affects another user's resources
- Ensure you have proper authorization
- Explain impact to target user's environment
` : `
USER OPERATION:
- Affects only current user's resources
- No special permissions needed
`}
STEP 6: COMMAND EXECUTION WITH ludus_cli_execute
Use the ludus_cli_execute MCP tool for all CLI command execution:
- DO NOT include "ludus" prefix (tool adds it automatically)
- Use: \`ludus_cli_execute({ command: "range status" })\`
- NOT: \`ludus_cli_execute({ command: "ludus range status" })\`
- For help: \`ludus_cli_execute({ command: "--help" })\` or \`ludus_cli_execute({ command: "range --help" })\`
- NOT: \`ludus_cli_execute({ command: "help" })\` (this subcommand doesn't exist)
${target_user ? `- Include user parameter: \`ludus_cli_execute({ command: "range status", user: "${target_user}" })\`` : ''}
STEP 7: ERROR HANDLING & TROUBLESHOOTING
If ludus_cli_execute command fails:
1. Check command syntax with help: \`ludus_cli_execute({ command: "command --help" })\`
2. Verify connectivity and permissions
3. Try simpler related commands to isolate the issue
4. Suggest specific tools if available for the task
CRITICAL BEHAVIORAL RULES:
SAFETY FIRST: Always prioritize user safety and data protection
CHECK BEFORE ACTING: Understand current state before making changes
COMMUNICATE CLEARLY: Explain what commands will do before execution
CONFIRM DESTRUCTIVE: Never assume destructive actions are wanted
USE SPECIFIC TOOLS: Prefer dedicated tools over raw CLI
ADMIN AWARENESS: Be explicit about admin operations affecting other users
USE CLI WRAPPER: Execute all CLI commands through ludus_cli_execute tool
EXECUTION GUIDANCE:
${isDestructiveConfirmed ?
'PROCEED with destructive action - user has confirmed' :
'If destructive: STOP and ask for confirmation first'
}
Execute the command now using the ludus_cli_execute tool following this protocol exactly.`
}
}
]
};
}
