# AI Loop Library Catalog

Last updated: 2026-07-01
Loop count: 43

AI Loop Library catalogs bounded, verifiable AI agent loops: trigger, objective, repeated action, verification, stop condition, budget, approval boundary, and safe output.

## Loops

### Source Library Ingestion QA
- URL: https://ailooplibrary.com/loops/source-library-ingestion-qa/
- Category: Knowledge
- Trigger: After each source capture
- Objective: Turn raw captures into retrieval-ready knowledge instead of a pile of orphaned links.
- Verification: Metadata complete, transcript/article state honest, useful takeaways present, and qmd retrieval verified or refreshed.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before public, destructive, financial, legal, account, or production-impacting actions.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Claude Code Repo Readiness
- URL: https://ailooplibrary.com/loops/claude-code-repo-readiness/
- Category: Engineering
- Trigger: Before major agent work
- Objective: Make the repo legible to agents before asking them to do real work.
- Verification: Repo has agent instructions, documented commands, architecture notes, risk areas, and a docs/loops scaffold.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before public, destructive, financial, legal, account, or production-impacting actions.
- Safe output: Pull request, patch, report, or evidence log
- Works with: Claude Code, OpenAI Codex, Cursor, Gemini CLI, any tool-using coding agent

### Project Docs Freshness
- URL: https://ailooplibrary.com/loops/project-docs-freshness/
- Category: Engineering
- Trigger: Nightly or after meaningful code changes
- Objective: Keep documentation synced with what the product actually does.
- Verification: Changed behavior, APIs, CLI commands, config, and workflows are reflected in docs. Docs checks pass.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before public, destructive, financial, legal, account, or production-impacting actions.
- Safe output: Pull request, patch, report, or evidence log
- Works with: Claude Code, OpenAI Codex, Cursor, Gemini CLI, any tool-using coding agent

### Test and Logging Coverage
- URL: https://ailooplibrary.com/loops/test-logging-coverage/
- Category: Engineering
- Trigger: Weekly or before release
- Objective: Raise the floor for every future loop by improving the evidence system.
- Verification: Critical flows have useful tests and structured logs for representative success and failure paths.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before public, destructive, financial, legal, account, or production-impacting actions.
- Safe output: Pull request, patch, report, or evidence log
- Works with: Claude Code, OpenAI Codex, Cursor, Gemini CLI, any tool-using coding agent

### CI Optimization
- URL: https://ailooplibrary.com/loops/ci-optimization/
- Category: Engineering
- Trigger: Monthly or when CI is painful
- Objective: Attack one CI bottleneck at a time until the target is met or the bottleneck report is more valuable than another tweak.
- Verification: CI p50/p95 improves against the same workflow without weakening tests or hiding failures.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before public, destructive, financial, legal, account, or production-impacting actions.
- Safe output: Pull request, patch, report, or evidence log
- Works with: Claude Code, OpenAI Codex, Cursor, Gemini CLI, any tool-using coding agent

### Open Loop and Stale Memory Cleanup
- URL: https://ailooplibrary.com/loops/open-loop-stale-memory-cleanup/
- Category: Knowledge
- Trigger: Weekly
- Objective: Reconcile what the system thinks is open against what actually happened.
- Verification: No current open loop is contradicted by recent daily or project notes.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before public, destructive, financial, legal, account, or production-impacting actions.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Social Source to Insight
- URL: https://ailooplibrary.com/loops/social-source-to-insight/
- Category: Content
- Trigger: After saving a high-signal source
- Objective: Convert high-signal sources into grounded ideas without auto-posting like a caffeinated intern.
- Verification: Source captured, takeaways extracted, draft angles written, and no public post is published without approval.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Fresh Clone Onboarding
- URL: https://ailooplibrary.com/loops/fresh-clone-onboarding/
- Category: Engineering
- Trigger: Before onboarding
- Objective: Prove the README works from zero, not tribal memory.
- Verification: A clean machine reaches the documented ready state using only the README.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Pull request, patch, report, or evidence log
- Works with: Claude Code, OpenAI Codex, Cursor, Gemini CLI, any tool-using coding agent

### Production Error Sweep
- URL: https://ailooplibrary.com/loops/production-error-sweep/
- Category: Operations
- Trigger: Daily or after incident
- Objective: Turn recurring production errors into verified patches.
- Verification: Actionable errors are fixed with reproduction or tests, or explicitly classified as noise.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Dependency CVE Burndown
- URL: https://ailooplibrary.com/loops/dependency-cve-burndown/
- Category: Security
- Trigger: After security scan
- Objective: Rank vulnerabilities by real exposure, then fix the reachable ones first.
- Verification: No exploitable high or critical CVE remains without an explicit risk decision.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Test Flake Stabilizer
- URL: https://ailooplibrary.com/loops/test-flake-stabilizer/
- Category: Engineering
- Trigger: When tests are inconsistent
- Objective: Find the real cause of flakes instead of wallpapering them with sleeps.
- Verification: The repaired test and full suite pass for the required consecutive-run streak.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Pull request, patch, report, or evidence log
- Works with: Claude Code, OpenAI Codex, Cursor, Gemini CLI, any tool-using coding agent

### Pre-Publish Source Check
- URL: https://ailooplibrary.com/loops/prepublish-source-check/
- Category: Content
- Trigger: Before publishing factual work
- Objective: Inventory the claims before the internet does it for you.
- Verification: Every checkable claim is supported by a current source or visibly flagged for an editor.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Research to Artifact
- URL: https://ailooplibrary.com/loops/research-to-artifact/
- Category: Knowledge
- Trigger: Whenever research must support a decision
- Objective: Constrain research around the decision it needs to support.
- Verification: The artifact meets acceptance criteria, traces important claims to sources, and states uncertainty plainly.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Browser Quality Streak
- URL: https://ailooplibrary.com/loops/browser-quality-streak/
- Category: Evaluation
- Trigger: Before release
- Objective: Run real scenarios until the streak proves the product is stable.
- Verification: N realistic scenarios pass consecutively, and earlier failures have regression coverage.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Error Message Rewrite
- URL: https://ailooplibrary.com/loops/error-message-rewrite/
- Category: Design
- Trigger: When users hit confusing errors
- Objective: Replace error sludge with clear recovery paths.
- Verification: Every in-scope user-visible error is accounted for, rewritten or blocked, and verified in a reachable state.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Accessibility Repair
- URL: https://ailooplibrary.com/loops/accessibility-repair/
- Category: Design
- Trigger: Before launch or when audits fail
- Objective: Fix access barriers without silencing checks or weakening the standard.
- Verification: No confirmed accessibility blocker remains in the agreed pages, components, or tasks.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Cold Load Trim
- URL: https://ailooplibrary.com/loops/cold-load-trim/
- Category: Engineering
- Trigger: When first visit feels heavy
- Objective: Reduce first-load bytes without breaking the first impression.
- Verification: Initial screen downloads fewer bytes while screenshots and behavior remain unchanged.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Pull request, patch, report, or evidence log
- Works with: Claude Code, OpenAI Codex, Cursor, Gemini CLI, any tool-using coding agent

### Architecture Rubric Refactor
- URL: https://ailooplibrary.com/loops/architecture-rubric-refactor/
- Category: Engineering
- Trigger: When architecture work has a defined scope
- Objective: Refactor against a written rubric, not taste fumes.
- Verification: Scoped module meets the written rubric, tests pass, and unresolved objections are explicit.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Pull request, patch, report, or evidence log
- Works with: Claude Code, OpenAI Codex, Cursor, Gemini CLI, any tool-using coding agent

### Adversarial PR Review
- URL: https://ailooplibrary.com/loops/adversarial-pr-review/
- Category: Evaluation
- Trigger: For meaningful PRs
- Objective: Separate builder and critic so “looks good to me” means something.
- Verification: An independent critic approves the unchanged version or only accepted findings remain.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Living Story
- URL: https://ailooplibrary.com/loops/living-story/
- Category: Operations
- Trigger: Weekly or per project window
- Objective: Maintain an evidence-backed story of what matters and what is unfinished.
- Verification: Every prior thread is carried forward, closed with evidence, or flagged stale/needs-review.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Promise to Proof
- URL: https://ailooplibrary.com/loops/promise-to-proof/
- Category: Growth
- Trigger: Before launch or campaign refresh
- Objective: Make product claims survive contact with reality.
- Verification: Every high-risk customer promise is supported, narrowed, or awaiting an explicit decision.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### SEO/GEO Visibility
- URL: https://ailooplibrary.com/loops/seo-geo-visibility/
- Category: Growth
- Trigger: Monthly
- Objective: Map priority questions to answer-ready pages.
- Verification: Priority pages are indexable, answer-ready, internally linked, and no high-impact benchmark gaps remain.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### One Post a Week Experiment
- URL: https://ailooplibrary.com/loops/one-post-a-week/
- Category: Content
- Trigger: Weekly for six weeks
- Objective: Change one variable per week and let useful audience behavior decide.
- Verification: Winner is supported by comparable replies, saves, and questions, not just likes.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Talk to Five Buyers
- URL: https://ailooplibrary.com/loops/talk-to-five-buyers/
- Category: Growth
- Trigger: Before major landing-page rewrite
- Objective: Let buyer language rewrite the page where the objection appears.
- Verification: Proposed copy traces to repeated anonymized buyer objections across approved interviews.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Agent Handoff Continuity
- URL: https://ailooplibrary.com/loops/agent-handoff-continuity/
- Category: Operations
- Trigger: Before pausing or switching agents
- Objective: Prevent the next agent from starting with amnesia and bravado.
- Verification: A new agent can state goal, current state, proof, blockers, and next action without reading the whole transcript.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Agent Toolchain Health Check
- URL: https://ailooplibrary.com/loops/toolchain-health-check/
- Category: Operations
- Trigger: Weekly or before a heavy agent run
- Objective: Check the harness before blaming the model.
- Verification: Critical tools authenticate, return sane output, and have a known fallback or blocker owner.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### API Contract Drift
- URL: https://ailooplibrary.com/loops/api-contract-drift/
- Category: Engineering
- Trigger: After API changes or SDK releases
- Objective: Find where the API contract forked and bring every consumer back to one truth.
- Verification: Server behavior, client types, examples, and docs agree on request/response contracts.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Pull request, patch, report, or evidence log
- Works with: Claude Code, OpenAI Codex, Cursor, Gemini CLI, any tool-using coding agent

### Calendar Promise Proof
- URL: https://ailooplibrary.com/loops/calendar-promise-proof/
- Category: Personal Ops
- Trigger: Daily or before week planning
- Objective: Mine the calendar for promises before they become social debt.
- Verification: Every calendar-derived promise has an owner, due date, evidence, or explicit drop decision.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Support Refund Follow-Up
- URL: https://ailooplibrary.com/loops/support-refund-followup/
- Category: Personal Ops
- Trigger: Until terminal state
- Objective: Keep the case moving until the money lands or the wall is real.
- Verification: Refund is received, or no approved next step remains and a blocker is named.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Loop Hiring Manager
- URL: https://ailooplibrary.com/loops/loop-hiring-manager/
- Category: Strategy
- Trigger: When a project accumulates recurring chores
- Objective: Hire loops like employees: for recurring work with clear outcomes.
- Verification: At most three candidates fill proven recurring gaps, with one recommended safe manual trial.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Acceptance Scenario Lockstep
- URL: https://ailooplibrary.com/loops/acceptance-scenario-lockstep/
- Category: Engineering
- Trigger: Before and during ambiguous feature work
- Objective: Pin the target with executable scenarios before the agent starts changing code.
- Verification: The same scenarios written before implementation pass after the change, and any scope expansion is explicitly approved.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Pull request, patch, report, or evidence log
- Works with: Claude Code, OpenAI Codex, Cursor, Gemini CLI, any tool-using coding agent

### Inbox Decision Triage
- URL: https://ailooplibrary.com/loops/inbox-decision-triage/
- Category: Personal Ops
- Trigger: Daily or before deep work
- Objective: Turn inbox sludge into a small list of decisions, drafts, and explicit no-actions.
- Verification: Every sampled message is archived, labeled, delegated, converted to a task, or left with a drafted reply; no external message is sent without approval.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Reference Oracle Implementation
- URL: https://ailooplibrary.com/loops/reference-oracle-implementation/
- Category: Engineering
- Trigger: Before implementing tricky behavior with an external source of truth
- Objective: Give the agent an oracle, not a pep talk.
- Verification: Generated outputs match the reference oracle across the agreed fixture set, with tolerances documented for legitimate differences.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Pull request, patch, report, or evidence log
- Works with: Claude Code, OpenAI Codex, Cursor, Gemini CLI, any tool-using coding agent

### Agent Instructions After-Action
- URL: https://ailooplibrary.com/loops/agent-instructions-after-action/
- Category: Engineering
- Trigger: After a successful or painful agent coding session
- Objective: Turn one agent run into better future agent runs.
- Verification: Repo instructions contain only reusable, source-grounded lessons and the next similar task can start without rediscovering the same trap.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before public, destructive, financial, legal, account, or production-impacting actions.
- Safe output: Pull request, patch, report, or evidence log
- Works with: Claude Code, OpenAI Codex, Cursor, Gemini CLI, any tool-using coding agent

### Spec to Task Shards
- URL: https://ailooplibrary.com/loops/spec-to-task-shards/
- Category: Engineering
- Trigger: Before multi-file feature work or migrations
- Objective: Make the agent split the work before it starts swinging a hammer.
- Verification: A written spec, non-goals, acceptance checks, and ordered task shards exist before implementation begins.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Pull request, patch, report, or evidence log
- Works with: Claude Code, OpenAI Codex, Cursor, Gemini CLI, any tool-using coding agent

### Behavior Ladder TDD
- URL: https://ailooplibrary.com/loops/behavior-ladder-tdd/
- Category: Engineering
- Trigger: When implementing logic-heavy features
- Objective: Use tests as the prompt, one behavior at a time.
- Verification: Each behavior test fails before implementation, passes after the smallest change, and remains green through final refactor.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before public, destructive, financial, legal, account, or production-impacting actions.
- Safe output: Pull request, patch, report, or evidence log
- Works with: Claude Code, OpenAI Codex, Cursor, Gemini CLI, any tool-using coding agent

### Trace-First Debugging
- URL: https://ailooplibrary.com/loops/trace-first-debugging/
- Category: Engineering
- Trigger: When an agent is tempted to patch a bug from a hunch
- Objective: Make the agent earn the fix before touching code.
- Verification: The bug is reproduced, the root cause is evidenced by traces or tests, and the fix includes a regression check.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before public, destructive, financial, legal, account, or production-impacting actions.
- Safe output: Pull request, patch, report, or evidence log
- Works with: Claude Code, OpenAI Codex, Cursor, Gemini CLI, any tool-using coding agent

### Visual Feedback Repair
- URL: https://ailooplibrary.com/loops/visual-feedback-repair/
- Category: Design
- Trigger: After AI-generated UI changes or visual bug reports
- Objective: Show the agent the page, not just the code.
- Verification: Before/after screenshots, console state, and responsive checks show the visual issue is fixed without breaking adjacent states.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Memory Bank Continuity
- URL: https://ailooplibrary.com/loops/memory-bank-continuity/
- Category: Knowledge
- Trigger: At the start and end of multi-session agent work
- Objective: Give stateless agents a project spine.
- Verification: Project brief, active context, decisions, tech constraints, and progress are current enough for a fresh agent to continue safely.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before public, destructive, financial, legal, account, or production-impacting actions.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Sandboxed YOLO Probe
- URL: https://ailooplibrary.com/loops/sandboxed-yolo-probe/
- Category: Security
- Trigger: Before allowing autonomous shell-heavy agent runs
- Objective: Let the agent run wild somewhere boring.
- Verification: The agent can run needed commands inside the sandbox, cannot reach forbidden files/secrets, and produces a replayable diff or report before host-side changes.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Parallel Agent Worktree Sweep
- URL: https://ailooplibrary.com/loops/parallel-agent-worktree-sweep/
- Category: Engineering
- Trigger: When several independent repo improvements can run at once
- Objective: Run more agents, but keep their blast radiuses separate.
- Verification: Each agent branch has isolated scope, passing checks, a summary, and no conflicting files before integration review.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Pull request, patch, report, or evidence log
- Works with: Claude Code, OpenAI Codex, Cursor, Gemini CLI, any tool-using coding agent

### Agent Merge Queue Review
- URL: https://ailooplibrary.com/loops/agent-merge-queue-review/
- Category: Evaluation
- Trigger: After multiple agent-generated PRs or branches accumulate
- Objective: Separate useful agent output from merge-shaped confetti.
- Verification: Only branches with passing checks, clear intent, non-conflicting scope, and human-readable evidence are merged or promoted.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Draft, report, checklist, table, or approval-gated recommendation
- Works with: Claude, ChatGPT, Gemini, any tool-using AI assistant

### Completion Promise Loop
- URL: https://ailooplibrary.com/loops/completion-promise-loop/
- Category: Engineering
- Trigger: For scoped implementation tasks where half-finished output is the main risk
- Objective: Make the agent promise completion against explicit evidence, not vibes and a cheerful summary.
- Verification: Every acceptance criterion is checked with tests, browser evidence, logs, screenshots, or a clear blocker report before the agent stops.
- Stop condition: Stop when the verifier passes, the budget is exhausted, no progress is made, a blocker appears, or approval is required.
- Budget: Set a time, turn, token, retry, file, or dollar cap before running the loop.
- Approval boundary: Human approval required before publishing, sending, deleting, spending, changing accounts, touching production, or making reputational/legal/financial commitments.
- Safe output: Pull request, patch, report, or evidence log
- Works with: Claude Code, OpenAI Codex, Cursor, Gemini CLI, any tool-using coding agent