How does the skill handle dependencies between prompts?

The skill automatically scans prompts for @ references and file paths to build a dependency graph, determining if tasks should be run sequentially or can be executed in parallel.

Can I run multiple prompts at the same time?

Yes, the skill features a parallel execution engine that can spawn multiple Task agents simultaneously for independent prompts that do not have shared dependencies.

What is a meta-prompt in this context?

A meta-prompt is a specialized prompt designed to generate output (like a research paper or a technical plan) specifically formatted for consumption by a subsequent Claude agent in a pipeline.

What are the quality gates mentioned in the description?

Quality gates are validation steps that check for specific XML tags in the AI's output, such as , , and , ensuring the output meets the required standard before proceeding.

Meta-Prompt Creator

Name: Meta-Prompt Creator
Author: marcusgoll

bymarcusgoll

•

Productivity & Workflow

Generates and orchestrates structured, multi-stage Claude-to-Claude pipelines for research, planning, and implementation tasks.

About

The create-meta-prompts skill enables Spec-Driven Development by breaking down complex product ideas into manageable, auditable prompt chains. It automates the creation of a standardized .prompts/ directory where tasks are divided into Research, Plan, and Do stages. By enforcing XML metadata like confidence scores and dependency tracking, the skill ensures that each step in an AI workflow provides high-quality input for the next, supporting sequential or parallel execution for reliable production-grade results.

Key Features

Flexible execution engine supporting sequential, parallel, and DAG-based modes
Intelligent dependency detection between prompt stages using @ references
Structured directory management with automated folder and file versioning
Quality gate enforcement via mandatory XML metadata validation
18 GitHub stars
Automated generation of multi-stage AI workflows (Research-Plan-Execute)

Use Cases

Architecting complex new features that require deep research before implementation
Creating auditable documentation and prompt chains for large-scale refactors
Automating repeatable AI workflows with built-in quality checks and token budgets

About

Key Features

Flexible execution engine supporting sequential, parallel, and DAG-based modes
Intelligent dependency detection between prompt stages using @ references
Structured directory management with automated folder and file versioning
Quality gate enforcement via mandatory XML metadata validation
18 GitHub stars
Automated generation of multi-stage AI workflows (Research-Plan-Execute)

Use Cases

Architecting complex new features that require deep research before implementation
Creating auditable documentation and prompt chains for large-scale refactors
Automating repeatable AI workflows with built-in quality checks and token budgets