If you find yourself writing the same kind of prompt over and over, or your team keeps reinventing workflows that should be standardized, this is the tooling layer for that problem. It helps you spot repeated work in your codebase or chat history, turn it into a bounded loop with clear success conditions, run it with an evidence receipt, and either save it to your project or submit it to a shared library. The interesting part is the audit mode: it checks loops for weak verification or runaway authority before you depend on them. Think of it as version control meets runbook discipline for agent patterns you actually want to repeat.
npx -y skills add forward-future/loop-library --skill loopy --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