using-superpowers

To ensure AI agents invoke relevant skills before generating responses, establish a mandatory instruction hierarchy that prioritizes skill execution over default system behavior for all conversational turns and task executions.

The instruction hierarchy for AI agent tool use prioritizes explicit user instructions first, followed by relevant skill execution, and finally default system prompts to prevent errors and ensure proper tool invocation.

Prioritize AI skills by requiring the agent to check and invoke any potentially relevant skill using the Skill tool before responding to user requests, overriding immediate default code generation or text output.

An AI agent skips tool use and generates immediate responses when a mandatory instruction hierarchy is missing, failing to enforce the rule that relevant skills must be considered and invoked before any other action.

Yes, you need to define an instruction hierarchy for simple AI workflow queries because mandatory skill invocation applies to all conversational turns, guaranteeing relevant tools are checked even for straightforward tasks.

The best way to structure AI agent workflows for instruction following is applying a consistent hierarchy across all task executions, ensuring the Skill tool is always checked before generating responses or taking actions.