create-meta-prompts

Generate XML-structured meta-prompts with provenance for Claude-to-Claude pipelines.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/caioniehues/dotfiles --skill create-meta-prompts-caioniehues
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: create-meta-prompts
Source: https://github.com/caioniehues/dotfiles/tree/main/skills/create-meta-prompts
Command: npx skills add https://github.com/caioniehues/dotfiles --skill create-meta-prompts-caioniehues

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill generates meta-prompts designed for Claude-to-Claude pipelines, enabling multi-stage workflows (research, plan, execute) with structured outputs and provenance. Each execution produces a SUMMARY.md for quick human scanning.

Core Features & Use Cases

  • Stage-aware prompts: Supports Research, Plan, Do, and Refine prompts with XML metadata.
  • Provenance and chaining: Automatically references prior prompts and maintains clear provenance.
  • SUMMARY.md generation: Produces executive summaries for quick human review.

Quick Start

Create a new meta-prompt using the /create-meta-prompt command to generate a chained research → plan → implement workflow and produce SUMMARY.md.

Frequently Asked Questions about create-meta-prompts

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate multi-stage Claude workflows with structured prompts?

Multi-stage Claude workflows use meta-prompts—templates that chain research, planning, and implementation stages with XML-structured outputs. This Skill automates their creation, generating dependency-aware prompts saved in organized folders with SUMMARY.md for each stage, enabling Claude-to-Claude pipelines with explicit provenance and automatic reference chaining.

What are meta-prompts and why use them for prompt engineering?

Meta-prompts are templates that guide Claude through structured multi-stage reasoning: research to gather context, planning to design solutions, and implementation to execute. They enforce consistent XML formatting, maintain provenance across stages, and enable dependency detection so each prompt automatically references prior artifacts, reducing manual coordination and improving workflow reliability.

How do I set up a research-to-implementation pipeline with Claude?

Use the /create-meta-prompt command to generate a chained workflow: a research prompt gathers information, a plan prompt structures the solution, and a do prompt executes it. The Skill creates all three with XML metadata, saves outputs to `.prompts/{number}-{topic}-{purpose}/`, generates SUMMARY.md for each stage, and automatically wires dependencies so each prompt references prior results.

Can I customize meta-prompts for specific project requirements?

Yes. Meta-prompts follow standardized patterns (do/plan/research/revise) and enforced schemas for folder structure and metadata, but each is generated for a specific topic and purpose. The Skill supports stage-aware customization—research, plan, and refine prompts adapt to your workflow while maintaining the XML structure and dependency chaining required for multi-stage execution.

What output does the meta-prompt generation produce?

Each execution produces XML-structured prompts organized in `.prompts/{number}-{topic}-{purpose}/`, a SUMMARY.md file for human review, and metadata encoding dependencies and provenance. Outputs are designed for Claude-to-Claude pipelines, so each downstream stage automatically receives prior research and planning artifacts without manual intervention.

Do I need prior prompt-engineering experience to use meta-prompts?

No. The Skill automates prompt creation and chaining logic, handling XML formatting, dependency detection, and folder schemas automatically. You define the topic and purpose; the Skill generates research, plan, and implementation prompts following proven patterns, eliminating the need to design multi-stage workflows from scratch or manually wire stage dependencies.