Which AI models can this skill execute prompts with?

The skill supports a diverse range of models including Claude subagents, OpenAI Codex (gpt-5.2), Google Gemini 2.5 Pro, Z.AI, and local models like Qwen and Devstral via specialized CLI tools.

Can I monitor prompts that are running in the background?

Yes, the skill includes a monitoring pattern that spawns a Haiku subagent to tail logs, check process IDs (PID), and report status every 30 seconds until the task is finished.

How does the verification loop feature work?

The verification loop allows the executor to run a task and check the output for a specific completion marker. If the task is incomplete or fails, it automatically retries the process until the 'VERIFICATION_COMPLETE' signal is detected.

Why would I use Git worktrees with prompt execution?

Worktrees allow you to run prompts in isolated directories. This prevents the AI model from conflicting with your current uncommitted changes and enables you to run multiple independent prompts in parallel safely.

AI Multi-Model Prompt Executor

Name: AI Multi-Model Prompt Executor
Author: cruzanstx

bycruzanstx

•

Productivity & Workflow

Executes repository-based prompts across multiple AI models with built-in worktree management and autonomous verification loops.

About

The AI Multi-Model Prompt Executor transforms Claude into a powerful orchestrator for complex development tasks. It allows users to execute prompts stored in a central directory using various LLMs like Codex and Gemini, managing the entire lifecycle from environment setup in isolated Git worktrees to background execution via tmux. With its advanced verification loop system, it can autonomously retry and correct tasks until they meet specific completion criteria, providing detailed monitoring and logging throughout the process to ensure high-quality, verified results.

Key Features

4 GitHub stars
Background execution monitoring with automated status updates and logging
Isolated Git worktree management for safe, parallel task execution
Multi-model CLI support for OpenAI Codex, Google Gemini, ZAI, and local models
Autonomous verification loops with self-correction and retry logic
Deep integration with tmux for persistent, long-running sub-tasks

Use Cases

Delegating specialized code generation or research to external high-reasoning models
Running multiple experimental feature implementations in parallel isolated environments
Automating complex, multi-step refactoring tasks that require iterative validation

About

Key Features

4 GitHub stars
Background execution monitoring with automated status updates and logging
Isolated Git worktree management for safe, parallel task execution
Multi-model CLI support for OpenAI Codex, Google Gemini, ZAI, and local models
Autonomous verification loops with self-correction and retry logic
Deep integration with tmux for persistent, long-running sub-tasks

Use Cases

Delegating specialized code generation or research to external high-reasoning models
Running multiple experimental feature implementations in parallel isolated environments
Automating complex, multi-step refactoring tasks that require iterative validation