Loop directory

Test challenger prompts on a working set, promote only on fresh holdout wins, and keep the champion when results are uncertain.

Run one agent in an isolated worktree and release its staged output only after a second agent verifies the work.

Open a PR, run an independent Codex review, fix every blocking finding, and repeat until it's clean.

A team-config pattern: a loop.md file in the project root overrides the built-in maintenance prompt, so any teammate running bare /loop gets your project's canonical loop — run lint, typecheck, and tests, fix anything red, stop when clean.

First-party Cursor guidance for the iterate-until-green loop, with the key anti-reward-hacking clause: the agent may never modify the tests it is trying to satisfy. Works in Cursor, Claude Code /goal, and Codex.

A hands-free ops loop that polls your deploy every two minutes, runs the smoke test the moment it goes live, and stops with a report if any check fails.

A drop-in Claude Code skill that keeps looping until an external verifier passes — not the model's own self-report — making it a strong anti-reward-hacking pattern for autonomous coding.

Profile the CI pipeline and cut wall-clock time under a target by improving caching, splitting slow jobs, and removing redundant steps, verifying every run stays green.

Move the project to the current Node LTS across .nvmrc, CI config, Dockerfiles, and engines, fixing deprecations until everything is green on the new runtime.

Rewrite the commit messages on your feature branch to conventional-commit format with meaningful bodies before opening the PR, leaving the code untouched.

Work a triaged bug list one fix per fresh-context iteration: reproduce first, fix minimally, prove it with a regression test, and log root-cause patterns to guardrails.

Document one undocumented public module per fresh-context iteration, verifying every code sample compiles and accumulating style rules in guardrails so the docs read like one author wrote them.

Break a large refactor into a JSON backlog of modules and let fresh-context iterations convert one module per pass, with guardrails capturing every pattern decision so the result stays consistent.

Iterate over a prioritized list of untested modules with fresh context each pass, writing real behavioral tests for one module at a time and banking lessons in a guardrails file.

The canonical Ralph loop: each iteration starts fresh, reads the PRD and guardrails, ships exactly one backlog item end-to-end, and records what it learned.

Every Monday, list remote branches with no commits in 30 days, classify each as merged, abandoned, or unclear, and file a cleanup checklist issue — deleting nothing.

Every Friday, compile a trend report of debt signals — TODO count, lint suppressions, type coverage, largest files — so the team sees drift before it compounds.

Run the full end-to-end suite every night at 2am and wake up to a digest of failures with screenshots, suspect commits, and a flakiness verdict per test.

Every Monday morning, run the dependency vulnerability audit, open one PR fixing what auto-fix can handle safely, and file issues for the rest.

Generate a contract test for every documented endpoint in the OpenAPI spec so the spec and the implementation can never silently drift.

Shrink the production JavaScript bundle under a hard budget by attacking the largest modules with code-splitting, lighter imports, and dead-weight removal.

Instrument the test suite with query logging, find endpoints issuing N+1 queries, and fix them with eager loading or batching until the hot paths are clean.

Walk through outdated dependencies one package per turn, upgrading, running the full check suite, and pinning back anything that breaks.

Hunt down unreferenced exports, unused files, and unreachable branches, deleting them in small verified steps until the analyzer reports clean.

Chip away at the oldest TODO comments in the codebase, one per pass: either resolve it properly or convert it into a tracked issue.

Keep a stack of feature branches rebased in order as their parents merge, resolving trivial conflicts and flagging risky ones for a human.

A Ralph-style loop that writes its own rules: when a check fails the same way twice, the failure pattern gets appended to a guardrails file that every later iteration reads first.

A loop wrapping a goal wrapping a review: every 30 minutes, drive all PR review comments to resolved via /review, 10 turns max per pass.

Run migrations, fix schema or SQL errors, and repeat until prisma migrate status reports clean, capped at 6 turns.

Review your recent diff for debug code, dead branches, and bad names, then fix with minimal edits until lint and tests pass.