How does this skill handle prompt dependencies?

The skill scans the .prompts/ directory for existing research and planning files, matching them by topic keywords to suggest relevant context for new execution prompts.

When should I use meta-prompts instead of standard prompts?

Use meta-prompts for complex, multi-step tasks that require planning or research. Avoid them for simple, single-step tasks or basic conversational queries where the overhead of folder creation isn't needed.

What is a Claude-to-Claude pipeline?

It is a workflow where the structured output of one Claude prompt (like a research document) serves as the direct context or input for a subsequent prompt (like an implementation plan), allowing for complex reasoning.

Can I run multiple prompts at the same time?

Yes, the skill includes a Parallel Execution engine that can spawn multiple Task agents simultaneously for independent sub-tasks, consolidating the results once complete.

AI Workflow Meta-Prompts

Name: AI Workflow Meta-Prompts
Author: Git-Fg

byGit-Fg

•

Productivity & Workflow

Automates the creation and execution of structured multi-stage Claude-to-Claude prompt pipelines for complex development tasks.

This skill enables the orchestration of sophisticated AI workflows by transforming high-level objectives into modular, multi-stage pipelines consisting of research, planning, and execution phases. It automatically manages a structured directory for prompts and outputs, detects dependencies between tasks, and provides execution engines for both sequential and parallel processing. By enforcing standardized XML structures and validation checks, it ensures that complex Claude-to-Claude interactions remain organized, reproducible, and highly effective for deep technical work.

Key Features

01Sequential and parallel execution engines for task agents

021 GitHub stars

03Automatic dependency and chain detection based on project context

04Multi-stage pipeline generation (Research, Plan, and Do phases)

05Structured .prompts directory management for workflow organization

06Automated validation and summary generation for every step

Use Cases

01Building recursive documentation systems where AI-generated plans drive implementation details

02Conducting deep-dive technical research before implementing new library integrations

03Orchestrating multi-file refactoring projects that require distinct planning and execution steps

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add git-fg/thecattoolkit generating-prompts

For use in Claude.ai and ChatGPT

Download Skill