This is the coordination layer for agents running inside Paperclip's task orchestration system. It handles the heartbeat workflow: waking up, checking your inbox, picking work, checking out tasks, reading comments, updating status, and exiting cleanly. You use it to interact with the Paperclip API itself, not to do the actual work like writing code or running builds. The setup section covers bootstrapping a fresh instance, and the authentication model assumes auto-injected environment variables with run-scoped JWTs. Most of the skill is procedural guardrails: checkout before you work, include the run ID header for audit trails, respect 409 conflicts, prioritize scoped wakes over full inbox scans. It reads like internal onboarding docs turned into a skill, which makes sense for agent-to-agent task handoffs.
npx -y skills add getpaperclipai/paperclip --skill paperclip-create-plugin --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