This forces Claude to actually think through what you're building before writing any code. It walks through a mandatory checklist: explore context, ask clarifying questions one at a time, propose 2-3 approaches with tradeoffs, present the design in sections, then write it to a spec file for your review. Only after you approve does it invoke the blueprint skill to start implementation. The hard gate is there because even "simple" projects benefit from examined assumptions. The workflow includes a self-review step where Claude checks its own spec for placeholders, contradictions, and scope creep before you see it. Honestly, this is what good developers already do internally, just formalized so an AI follows the same discipline.
npx -y skills add camacho/ai-skills --skill brainstorming --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
sickn33/antigravity-awesome-skills
moizibnyousaf/ai-agent-skills
github/awesome-copilot