Practical patterns, starters & CLI tools for loop engineering with AI coding agents. Design systems that prompt and orchestrate agents (inspired by Addy Osmani and Boris Cherny). Includes loop-audit, loop-init, loop-cost.
TypeScript agentic-aiai-agentsai-codinganthropicautomationclaudeclaude-codecodexcoding-agentsdevops-automationdevtoolsgithub-actionsgrokllmloop-engineeringmcpprompt-engineering
Post-Merge Cleanup — OpenClaw
Follow-up tech debt after merges land on main. Low-risk L1 loop — pairs naturally with Daily Triage.
Prerequisites
mkdir -p skills/post-merge-scan
cp templates/SKILL.md.loop-triage skills/post-merge-scan/SKILL.md
cp starters/minimal-loop/STATE.md.example post-merge-state.md
Report-Only (Week 1)
Run off-peak (evening) to avoid colliding with active work:
openclaw cron create "0 2 * * *" \
--name "Post-merge cleanup" \
--session isolated \
--message "Run skills/post-merge-scan/SKILL.md. Read post-merge-state.md first. Scan merges to main in the last 48h. Propose small doc/lint/comment fixes only — ticket anything architectural or multi-file. Update post-merge-state.md. Do not edit source code." \
--tools read
With Small Auto-Fixes (Week 2+)
openclaw cron edit <job-id> \
--message "Run post-merge-scan. For the highest-priority single-file cleanup: implement the fix in a worktree, run tests, have a verifier review. Escalate to a human for anything on denylist paths."
Safety
- Run off-peak to avoid collision with active branches.
- Never auto-fix architectural debt — file a ticket.
- Denylist paths always go to a human.
- Max 2 fix attempts per run.