atomic-blocks

Compose 16 atomic blocks for categorical meta-prompting workflows.

1|Updated Feb 2, 2026

Search Tags:#AI workflow #composition #category theory #meta-prompting #prompting #LLM primitives

Ready to run

npx skills add https://github.com/HermeticOrmus/hermetic-claude --skill atomic-blocks-hermeticormus

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: atomic-blocks
Source: https://github.com/HermeticOrmus/hermetic-claude/tree/main/claude/skills/atomic-blocks
Command: npx skills add https://github.com/HermeticOrmus/hermetic-claude --skill atomic-blocks-hermeticormus

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill provides a foundational set of composable "atomic blocks" that enable sophisticated and flexible meta-prompting strategies, moving beyond simple prompts to structured, verifiable, and adaptable AI workflows.

Core Features & Use Cases

Modular Design: Defines 16 reusable blocks across Assessment, Transformation, Refinement, and Composition layers.
Progressive Disclosure: Supports simple commands for 90% of users, while allowing power users to override or compose blocks for advanced control.
Use Case: A developer can use /compose with specific blocks to create a highly customized code review process that prioritizes security assessments, followed by performance checks, and then documentation generation, all orchestrated by the atomic blocks.

Quick Start

Use the atomic-blocks skill to assess the difficulty of the task "implement a caching layer".

Frequently Asked Questions about atomic-blocks

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

FAQPage Schema

What is categorical meta-prompting and how does it structure AI workflows?▼

Categorical meta-prompting structures AI workflows using composable primitives across assessment, transformation, refinement, and composition layers. This framework enables progressive disclosure, allowing simple commands for basic use and block overrides for advanced control.

How do I build a custom code review workflow using composable LLM primitives?▼

You build a custom code review workflow by using composition commands to chain specific atomic blocks. This orchestrates targeted processes like security assessments, performance checks, and documentation generation into a single verifiable AI workflow.

Do I need to understand category theory to use these meta-prompting blocks?▼

You do not need to understand category theory for standard usage. The framework supports progressive disclosure, providing simple commands for most users, while power users can leverage categorical and monadic principles for custom block overrides.

What's the best way to create adaptable prompt chains instead of static prompts?▼

The best way to create adaptable prompt chains is utilizing modular atomic blocks. These 16 reusable components allow you to construct flexible meta-prompting strategies that can be explicitly composed and overridden for varying task requirements.

Can I override individual blocks within an AI workflow composition?▼

Yes, you can override individual blocks within an AI workflow composition. The framework provides 16 reusable atomic blocks that support explicit composition and custom overrides, ensuring mathematical coherence and testability for advanced control.

Why does my meta-prompting workflow require adherence to monadic principles?▼

Adherence to monadic principles ensures mathematical coherence and testability within your meta-prompting workflow. This categorical foundation guarantees that composed atomic blocks function predictably across the assessment, transformation, refinement, and composition layers.