When you ship AI-generated code, you're handing off intent that was never written down. This skill defines the minimal doc set that makes vibe-coded apps reviewable: architecture, user flows with authz checks at every trust boundary, permissions, secrets mapped to rotation risk, and a brutally honest test coverage map that separates what's verified today from what's only proposed. Conditional docs for email, cron, SEO, and embedded agents get added only when those surfaces exist. The output lives in `/documentation` and serves two readers: the human reviewer and the next agent that touches the codebase. It's reverse-engineered from code, not aspirational, so audits have something real to compare against.
npx -y skills add phuryn/pm-skills --skill shipping-artifacts --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
juliusbrussee/caveman
mattpocock/skills
obra/superpowers
forrestchang/andrej-karpathy-skills
vercel-labs/skills