Make any repeated or self-paced agent work safe — a written goal, a machine-checkable stop condition, a cap on attempts and time, a report after every round, and a clean way to interrupt.
Install
npx skills add gaurav8936/oneroof --skill bounded-loopsnpx skills add gaurav8936/oneroof --skill bounded-loops -a claude-code -gnpx skills add gaurav8936/oneroof --skill bounded-loops -a codex -ggemini skills install https://github.com/gaurav8936/oneroof.git --path agentic_engineering/skills/fleet/bounded-loopsgh skill install gaurav8936/oneroof bounded-loopsgit clone --depth 1 https://github.com/gaurav8936/oneroof /tmp/gaurav8936__oneroof
mkdir -p ~/.claude/skills ~/.agents/skills
cp -R /tmp/gaurav8936__oneroof/agentic_engineering/skills/fleet/bounded-loops ~/.claude/skills/bounded-loops # Claude Code
cp -R /tmp/gaurav8936__oneroof/agentic_engineering/skills/fleet/bounded-loops ~/.agents/skills/bounded-loops # Codex, Cursor, Gemini CLI, Copilot, OpenCode