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
Dependency Sweeper — OpenClaw
Discover, safely apply, and verify dependency updates. OpenClaw's worktree isolation via exec + git worktree keeps each attempt safe.
Prerequisites
mkdir -p skills/dependency-triage skills/loop-verifier skills/minimal-fix
cp templates/SKILL.md.loop-triage skills/dependency-triage/SKILL.md
cp templates/SKILL.md.verifier skills/loop-verifier/SKILL.md
cp templates/SKILL.md.minimal-fix skills/minimal-fix/SKILL.md
cp starters/minimal-loop/STATE.md.example dependency-sweeper-state.md
L1 — Report Mode (Week 1)
openclaw cron create "0 */6 * * *" \
--name "Dependency sweeper" \
--session isolated \
--message "Run skills/dependency-triage/SKILL.md. Scan package manifests and lockfiles for outdated and vulnerable deps. Report findings only — do not edit any files. Update dependency-sweeper-state.md with the top-priority items. Escalate majors and high-sev CVEs." \
--tools read
L2 — Patch-Only Auto-Fix (Week 3+)
openclaw cron edit <job-id> \
--message "Run dependency-triage. For the highest-severity patch-only upgrade: apply the fix in a worktree, run npm ci && npm test, have a verifier review the diff. Never touch denylisted packages. Escalate majors to a human."
Verification Split
| Role | OpenClaw shape |
|---|---|
| Triage | Isolated cron with --tools read |
| Implementer | Main agent or coding-agent in worktree |
| Verifier | Second agent id or checker block in cron --message |
Safety
- Patch-only by default. Majors and breaking changes escalate to human.
- Maintain a denylist in
dependency-sweeper-state.md. - Max 3 fix attempts per run; pause after 5 auto-PRs per day.