create-meta-prompts

Create optimized prompts for Claude-to-Claude pipelines with XML metadata.

15|3|Updated Oct 19, 2025
One-click install
npx skills add https://github.com/o2alexanderfedin/ai-swe-template --skill create-meta-prompts-o2alexanderfedin
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: create-meta-prompts
Source: https://github.com/o2alexanderfedin/ai-swe-template/tree/main/.claude/skills/create-meta-prompts/skills/create-meta-prompts
Command: npx skills add https://github.com/o2alexanderfedin/ai-swe-template --skill create-meta-prompts-o2alexanderfedin

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Create optimized prompts for Claude-to-Claude pipelines with research, planning, and execution stages. Use when building prompts that produce outputs for other prompts to consume, or when running multi-stage workflows (research -> plan -> implement).

Core Features & Use Cases

  • Establish a repeatable workflow for multi-stage prompting (research, plan, implement)
  • Organize outputs into structured artifacts (research.md, plan.md) with XML metadata
  • Enable deterministic, dependency-aware execution via a centralized prompt engine
  • Maintain clear provenance by placing each prompt in a dedicated folder (e.g., .prompts/{number}-{topic}-{purpose}/)
  • Support chain detection and reference-aware prompts to streamline complex workflows

Quick Start

Describe the purpose (Do/Plan/Research) and the topic to begin, then follow the adaptive intake flow to generate prompts and artifacts.

Frequently Asked Questions about create-meta-prompts

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

FAQPage Schema
How do I organize multi-stage prompts for Claude-to-Claude workflows?

Multi-stage prompting organizes research, planning, and implementation stages into structured folders with versioned artifacts and XML metadata. This Skill enforces a repeatable folder structure (`.prompts/{number}-{topic}-{purpose}/`) and standardized templates, enabling downstream prompts to consume outputs deterministically with clear provenance and dependency tracking.

What's the best way to create prompts that feed into other prompts?

Design prompts for Claude-to-Claude pipelines by establishing reference chains and dependency awareness. This Skill generates optimized prompts with XML metadata schemas and automatic provenance, organizing artifacts (research.md, plan.md) so each downstream prompt knows its inputs and can execute deterministically without manual handoff.

How do I maintain provenance and dependencies across prompt stages?

Provenance is maintained by placing each prompt in a dedicated versioned folder with XML metadata that tracks references and dependencies. This Skill automates dependency handling and chain detection, creating a centralized record of which prompts consumed which artifacts, enabling audit trails and dependency-aware re-execution.

Can I automate research, planning, and implementation workflows with prompts?

Yes. This Skill enables deterministic, dependency-aware execution of multi-stage workflows via a centralized prompt engine. It creates optimized prompts for each stage (research → plan → implement) with standardized templates and integration points, automating handoffs and eliminating manual coordination between stages.

When should I use XML metadata in prompt workflows?

Use XML metadata when running multi-stage workflows where downstream prompts must consume structured outputs from earlier stages. Metadata schemas enforce contract clarity between prompt stages, enable reference chaining, support dependency detection, and provide integration hooks for execution engines to orchestrate pipelines reliably.

Do I need a prompt execution engine to use multi-stage workflows?

An execution engine enables deterministic orchestration and dependency resolution across stages, but isn't strictly required. This Skill generates prompts and artifacts compatible with centralized engines; you can also run stages manually, though automation requires an engine to handle reference resolution and provenance tracking.