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
AGENTS.md — CI Sweeper (Opencode)
Loop Mode
- CI sweeper is L2 from week one (action-oriented — CIs are urgent).
- Read
ci-sweeper-state.mdbefore every run. - Classify failures before attempting any fix.
Safety
- Always dispatch into a git worktree per attempt.
- Infra / security / payments test failures: do not auto-fix; flag and stop.
- Max 3 fix attempts per item; escalate after.
- Enforce budget cap from
LOOP.md.
Verification
- Verifier must run tests after implementer.
- Record attempt count in
ci-sweeper-state.md. - No auto-merge without verifier approval.