Arming Thought

This is a lightweight routing framework that runs at the start of each top-level conversation to establish a "seek truth from facts" baseline before deciding whether to load other skills. It's basically a dispatch layer: you verify claims against actual evidence, check your work before calling it done, and only pull in downstream skills like contradiction-analysis or practice-cognition when the task genuinely needs them. The core idea is to avoid both empty speculation and mechanical skill-stacking. If you're already in a sub-agent handling a narrow task, you skip it entirely. Think of it as a sanity check that asks "do I actually need a framework here, or can I just do the work?"

Install to Claude Code

npx -y skills add hughyau/qiushi-skill --skill arming-thought --agent claude-code