loop-engineering
Design self-running agent loops that keep codebases healthy
All Skills in This Repository (21)
Pure Emerald Level Indicatorsgoal-verifier
Automate goal completion verification against defined criteria in project management.
pr-review-triage
Monitor pull requests for status, merge conflicts, and unanswered comments.
loop-triage
Automates triage of development changes, CI failures, and issues into prioritized reports using Python.
post-merge-scan
Identify technical debt, TODOs, and debugging code in recently merged code.
issue-triage
Scan open issues for duplicates, prioritize, label, and escalate critical ones.
ci-triage
Parse CI failure output and classify issues as flake, regression, environment, or configuration.
loop-verifier
Verify loop-produced code changes by running tests and validating scope.
changelog-scan
Extract PR and commit information into a structured release notes draft.
draft-release-notes
Convert structured change log entries into formatted release note drafts.
dependency-triage
Scan package manifests and lockfiles for outdated and vulnerable dependencies.
loop-constraints
Enforce system constraints during multi-step workflow executions.
loop-budget
Monitor token budget usage during loop runs and exit on breach.
Frequently Asked Questions
FAQPage SchemaHow to install loop-engineering?โผ
Run `npx skills add cobusgreyling/loop-engineering --all -g -y` in your terminal to install all loop skills globally.
What is loop engineering?โผ
It is a pattern library for running AI agents in scheduled loops around a codebase, so the system discovers work, hands it to agents, verifies results, and saves state instead of you typing the next prompt.
Which repo tasks can agent loops automate?โผ
Ready-made patterns cover daily triage, PR babysitting, CI failure sweeping, dependency updates, issue labeling, changelog drafting, and post-merge cleanup.
Is it safe to let loops change code automatically?โผ
Yes, by design every loop starts in report-only mode and only graduates to assisted or unattended fixes after a verifier agent has proven reliable for a week.
Does loop-engineering work with Claude Code and Codex?โผ
Yes. Starters and examples are included for Claude Code, Codex, Grok, Opencode, OpenClaw, and plain GitHub Actions.
Related Repositories in Software Engineering
View All in Software Engineeringโopenclaw
Run a personal AI assistant across your devices and chat apps
superpowers
Gives coding agents a disciplined workflow from idea to merged code
react
AI agent skills for building, testing, and porting React core