This gives Claude persistent memory for each of your projects by running deterministic Node.js scripts instead of relying on the LLM to remember context. You init a project once, then /ck:save after each session to log what you did, what's next, and any decisions made. When you return, /ck:resume loads everything back in. The session start hook auto-injects context (about 100 tokens) and warns you if there's been git activity since your last save. It's overkill if you're just tinkering, but if you're juggling multiple codebases or returning to projects after weeks away, having a structured state log beats trying to reconstruct what you were doing from commit messages. The v2 format stores everything in JSON instead of markdown, so it's machine-readable if you want to build on it.
npx -y skills add affaan-m/everything-claude-code --skill ck --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
juliusbrussee/caveman
mattpocock/skills
shadcn/improve
obra/superpowers
forrestchang/andrej-karpathy-skills
vercel-labs/skills