using-superpowers

Enforce skill invocation before responses with a defined instruction hierarchy.

Updated Aug 27, 2026

Search Tags:#workflow automation #prompt engineering #llm #ai agent #skill invocation #instruction following

Ready to run

npx skills add https://github.com/maddhruv/cuttlefish --skill using-superpowers-maddhruv

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: using-superpowers
Source: https://github.com/maddhruv/cuttlefish/tree/main/.agents/skills/using-superpowers
Command: npx skills add https://github.com/maddhruv/cuttlefish --skill using-superpowers-maddhruv

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill ensures AI agents correctly prioritize and invoke other skills or user instructions, preventing errors and ensuring efficient task execution.

Core Features & Use Cases

Instruction Hierarchy: Clearly defines the order of precedence between user instructions, skills, and default system prompts.
Skill Invocation Mandate: Enforces the use of relevant skills before any other action or response, even for simple queries.
Use Case: When a user asks a question that could be answered by a skill, this Skill ensures the AI first checks for and invokes the relevant skill, rather than answering directly or asking a clarifying question.

Quick Start

Always invoke any potentially relevant skill before responding to the user.

Frequently Asked Questions about using-superpowers

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

FAQPage Schema

How do I enforce strict instruction hierarchy for AI agents?▼

To enforce strict instruction hierarchy for AI agents, implement a protocol that prioritizes user commands over skills and default system prompts. This prevents agent errors by ensuring explicit user instructions always take precedence over automated skill logic.

Why does my LLM agent answer directly instead of invoking a relevant skill?▼

An LLM agent answers directly instead of invoking a relevant skill when a strict skill invocation mandate is missing. You must enforce a protocol requiring the agent to check for and execute relevant skills before generating any response or taking action.

How do I manage platform-specific tool mappings for Claude Code and Gemini CLI?▼

To manage platform-specific tool mappings for Claude Code and Gemini CLI, configure an automation protocol that translates standard skill invocations into the correct environment-specific tool actions. This ensures consistent workflow execution across different platforms.

What is the correct precedence order for user commands versus AI skills?▼

The correct precedence order for user commands versus AI skills places user instructions first, followed by skills, and then default system prompts. This hierarchy ensures user intent always overrides automated behaviors and system defaults.

Does workflow automation require checking for relevant skills before responding to simple queries?▼

Workflow automation for AI agents requires checking for relevant skills before responding to simple queries. This mandatory pre-check step ensures all actions leverage available specialized instructions, preventing direct answers when a relevant skill exists.