Design a reliable agent loop from a plain description of the work. Picks the loop shape, defines what one iteration may touch, sets the evidence each success criterion needs, writes a stop condition that measures progress instead of counting turns, and emits a LOOP.md contract plus the Claude Code configuration and verification hook that enforce it.
Install
npx skills add rhawk117/rygentic-harness --skill build-a-loopnpx skills add rhawk117/rygentic-harness --skill build-a-loop -a claude-code -gnpx skills add rhawk117/rygentic-harness --skill build-a-loop -a codex -ggemini skills install https://github.com/rhawk117/rygentic-harness.git --path plugins/vibe-code/skills/build-a-loopgh skill install rhawk117/rygentic-harness build-a-loopgit clone --depth 1 https://github.com/rhawk117/rygentic-harness /tmp/rhawk117__rygentic-harness
mkdir -p ~/.claude/skills ~/.agents/skills
cp -R /tmp/rhawk117__rygentic-harness/plugins/vibe-code/skills/build-a-loop ~/.claude/skills/build-a-loop # Claude Code
cp -R /tmp/rhawk117__rygentic-harness/plugins/vibe-code/skills/build-a-loop ~/.agents/skills/build-a-loop # Codex, Cursor, Gemini CLI, Copilot, OpenCode