Skills
.Work. You
Rest

metaprompt

Community

Principled metaprompting for any model.

Software Engineering#LLM#prompting#metaprompt#model-selection#tool-schemas#context-budgeting
Authorjeff-hamm
Version1.0.0
Installs0

System Documentation

What problem does it solve?

Corpus-agnostic prompting principles that help teams design prompts, tool schemas, and prompt-driven app behavior that work across Claude/OpenAI/Gemini-class APIs.

Core Features & Use Cases

  • Prompt structure guidance: system vs user vs assistant, prefilling, example ordering, and format specification.
  • Context + token budgeting: model limits, partitioning for reuse, and cache-aware design.
  • Model selection guidance: when to use Haiku, Sonnet, or Opus based on task class.
  • Tool-use schemas: naming conventions, input schemas, and safety hints (MCP annotations).
  • Progressive disclosure: catalog slots are loaded lazily with references surfaced as needed.

Quick Start

Define a stable system prompt and a lazily-loaded prompt catalog following these metaprompt rules.

Dependency Matrix

Required Modules

None required

Components

Standard package

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: metaprompt
Download link: https://github.com/jeff-hamm/ai-tools/archive/main.zip#metaprompt

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.