cc-thinking-skills

Why we included this project

Coding agents like Claude Code and Cursor are good at producing code, but they often need a reasoning structure when the task is ambiguous rather than mechanical. This skill set ships 28 Agent Skills that each hand the assistant an explicit procedure: scientific method to localize a bug, pre-mortem to stress-test a launch plan, theory of constraints to find a bottleneck, or reversibility to classify an architecture decision. You install them through the skills CLI or as a Claude Code plugin, then invoke a skill by name or let the router pick the right frame when you are unsure which fits. There is also a routing and composition layer for chaining skills together, and an evidence audit worth reading before you trust any performance claim. For teams that want their coding assistant to think through diagnosis, risk, and strategy rather than just produce output, this gives them something concrete to adopt.