create-meta-prompts

Create XML-structured metadata prompts for Claude-to-Claude multi-stage workflows.

Updated Dec 16, 2025

Search Tags:#planning #research #workflow automation #prompt engineering #meta-prompting #claude-to-claude #multi-stage

Ready to run

npx skills add https://github.com/Eric-Ness/AI_Notes --skill create-meta-prompts-eric-ness

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: create-meta-prompts
Source: https://github.com/Eric-Ness/AI_Notes/tree/main/.claude/skills/create-meta-prompts
Command: npx skills add https://github.com/Eric-Ness/AI_Notes --skill create-meta-prompts-eric-ness

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill eliminates the complexity of building Claude-to-Claude pipelines by automatically generating optimized prompts for research, planning, and execution stages.

Core Features & Use Cases

Multi-Stage Workflows: Automatically chain research → plan → implement prompts with dependency detection.
Structured Outputs: Create prompts that produce XML-structured outputs for easy parsing by subsequent prompts.
Use Case: Imagine you need to research authentication options, then plan the implementation, then build it - all with automatic context passing between stages.

Quick Start

Use the create-meta-prompt command to research Node.js authentication libraries, then automatically create follow-up planning and implementation prompts.

Core Features & Use Cases

Automatic Dependency Detection: Scans for existing research and plan files to reference.

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 workflows between Claude prompts?▼

Multi-stage Claude-to-Claude workflows chain research, planning, and implementation prompts with automatic dependency detection. This Skill generates metadata-rich prompts that output XML-structured results, SUMMARY.md files, and dependency metadata so each stage passes context reliably to the next without manual reformatting.

What's the best way to structure outputs so Claude can parse them in a follow-up prompt?▼

Structured outputs use XML and metadata artifacts—dependencies, open questions, and SUMMARY files—organized per-prompt in .prompts folders. This format enables reliable parsing and context chaining across multiple prompts, eliminating ambiguity and reformatting friction between stages.

Can I automatically detect dependencies between research, planning, and implementation tasks?▼

Yes. Automatic dependency detection scans existing research and plan files to reference them in follow-up prompts. The Skill generates prompts that enforce explicit dependency tracking and traceability, so downstream stages know exactly which upstream outputs they depend on.

How do I ensure provenance and traceability across chained Claude prompts?▼

Provenance tracking is built into the generated prompts through XML metadata, explicit SUMMARY.md generation, and per-prompt folder organization. Each prompt documents its dependencies, outputs, and questions, creating a complete audit trail of how results flowed through your multi-stage workflow.

Do I need to manually reformat research outputs before passing them to a planning prompt?▼

No. The Skill generates prompts that output XML-structured data and SUMMARY artifacts designed for direct consumption by the next stage. Dependency detection and structured metadata eliminate manual reformatting, so planning prompts can reference and parse research outputs automatically.