shinjaehyun20

package-authoring

Use this skill whenever adding, revising, or publishing an AI Workflow Kits package, including Codex skills, Claude agents, Gemini prompts, Copilot instructions, hooks, commands, plugins, examples, manifests, and registry entries. This skill keeps contributions organized by workflow package first, runtime second, and artifact type third, and it requires the public safety scan before publishing.

teamwork_preview

Build a worker-ready launch packet before multi-agent execution. Use when a large task needs goal lock, grill-me review, branch planning, objective acceptance criteria, explicit approval, and evidence-first fan-in.

github-publication-bundle

Use this skill when releasing a user-authored public workflow package and you need readiness checks, package docs, examples, verification, and GitHub-facing source-of-truth updates to move together instead of stopping at commit/push.

keepworking

Use this skill when the user wants persistent progress instead of a shallow answer: "keep working", "continue until verified", "debug until the cause is clear", "repair and re-run", "parallelize these checks", "do not stop at analysis", or any task that needs goal locking, tiered routing, evidence, repair, and re-verification before closure.

pet-companion

Use this skill when the user wants to package a Codex pet success case as a public-safe cross-runtime workflow, export a runtime-neutral bundle, add a small companion viewer, or document adapter guidance for Claude Code, GitHub Copilot, OpenClaw, or Paperclip.