Source Library Ingestion QA ★
Turn raw captures into retrieval-ready knowledge instead of a pile of orphaned links.
A new article, video, X post, podcast, or image enters your knowledge base and should be findable later.
Sortable wall
The loop library.
All 68 reusable loops in one searchable wall. Filter by category, difficulty, tag, or Paul originals when you already know the kind of work you want to run.
Loop Library
Search the full wall of copyable loops.
68 loops. Search title, use case, tags, verification, or prompt.
★ Paul's picks
Claude Code Repo Readiness ★
Make the repo legible to agents before asking them to do real work.
You are about to let a coding agent do meaningful work in a repo and want fewer dumb mistakes.
Project Docs Freshness ★
Keep documentation synced with what the product actually does.
Code changed and the docs are probably now lying by omission.
Test and Logging Coverage ★
Raise the floor for every future loop by improving the evidence system.
Important flows are hard to debug or verify, especially auth, payments, data mutation, external APIs, jobs, and error boundaries.
CI Optimization ★
Attack one CI bottleneck at a time until the target is met or the bottleneck report is more valuable than another tweak.
The team is waiting on CI long enough that it changes behavior.
Open Loop and Stale Memory Cleanup ★
Reconcile what the system thinks is open against what actually happened.
Your personal operating system has accumulated stale promises, half-decisions, or outdated context.
Social Source to Insight ★
Convert high-signal sources into grounded ideas without auto-posting like a caffeinated intern.
A strong X or YouTube source should become reusable insight and possible social material.
Fresh Clone Onboarding
Prove the README works from zero, not tribal memory.
You suspect setup instructions only work for people who already know the system.
Production Error Sweep
Turn recurring production errors into verified patches.
Production telemetry should become repairs, not vibes.
Dependency CVE Burndown
Rank vulnerabilities by real exposure, then fix the reachable ones first.
Scanners are noisy and you need to know what is actually reachable.
Test Flake Stabilizer
Find the real cause of flakes instead of wallpapering them with sleeps.
CI fails differently across comparable runs.
Pre-Publish Source Check
Inventory the claims before the internet does it for you.
A factual draft is about to go public.
Research to Artifact
Constrain research around the decision it needs to support.
Research should end in a decision-ready memo, not a heap of notes.
Browser Quality Streak
Run real scenarios until the streak proves the product is stable.
You need confidence that the product works in realistic flows.
Error Message Rewrite
Replace error sludge with clear recovery paths.
Your product exposes raw provider errors, stack traces, or dead-end messages.
Accessibility Repair
Fix access barriers without silencing checks or weakening the standard.
A site or app needs a defined accessibility target and repeatable testing.
Cold Load Trim
Reduce first-load bytes without breaking the first impression.
A web app ships too much before showing anything useful.
Architecture Rubric Refactor
Refactor against a written rubric, not taste fumes.
A refactor matters, but “make architecture better” is too vague.
Adversarial PR Review
Separate builder and critic so “looks good to me” means something.
One model built the change and should not grade its own homework.
Living Story
Maintain an evidence-backed story of what matters and what is unfinished.
Work spans repos, notes, goals, and people.
Promise to Proof
Make product claims survive contact with reality.
Marketing, docs, demos, and support answers may promise more than the product can prove.
SEO/GEO Visibility
Map priority questions to answer-ready pages.
You need search engines and AI answer engines to understand the site.
One Post a Week Experiment
Change one variable per week and let useful audience behavior decide.
You want a repeatable content format and need audience evidence.
Talk to Five Buyers
Let buyer language rewrite the page where the objection appears.
Landing-page copy is missing the concern that almost stopped buyers.
Agent Handoff Continuity
Prevent the next agent from starting with amnesia and bravado.
Long-running agent work crosses sessions, tools, repos, or humans.
Agent Toolchain Health Check
Check the harness before blaming the model.
Agent setup depends on CLIs, MCP servers, browsers, tokens, local models, and cron jobs.
API Contract Drift
Find where the API contract forked and bring every consumer back to one truth.
Backend, frontend, docs, and SDKs may describe different realities.
Calendar Promise Proof
Mine the calendar for promises before they become social debt.
Meetings create obligations that never make it into a task system.
Support Refund Follow-Up
Keep the case moving until the money lands or the wall is real.
A refund or support claim will take multiple conversations and deadlines.
Loop Hiring Manager
Hire loops like employees: for recurring work with clear outcomes.
You have lots of automation ideas and need to decide which loop deserves oxygen.
Acceptance Scenario Lockstep
Pin the target with executable scenarios before the agent starts changing code.
A coding agent is about to implement a feature where success could be interpreted three different ways.
Inbox Decision Triage
Turn inbox sludge into a small list of decisions, drafts, and explicit no-actions.
Email, DMs, or comments are creating invisible obligations and context switching.
Reference Oracle Implementation
Give the agent an oracle, not a pep talk.
You need a coding agent to implement complex behavior that can be checked against a browser, official library, legacy system, API, compiler, or production output.
Agent Instructions After-Action
Turn one agent run into better future agent runs.
A coding agent just learned something about the repo, test harness, edge case, command, or convention that future sessions should not relearn the expensive way.
Spec to Task Shards
Make the agent split the work before it starts swinging a hammer.
The request is too large or vague for a single coding-agent prompt but not large enough to justify a full project plan ceremony.
Behavior Ladder TDD
Use tests as the prompt, one behavior at a time.
A pricing rule, validator, parser, permissions rule, workflow state machine, or other logic-heavy feature could balloon if implemented all at once.
Trace-First Debugging
Make the agent earn the fix before touching code.
A failure has logs, screenshots, stack traces, recent commits, or user reports, but the cause is not obvious yet.
Visual Feedback Repair
Show the agent the page, not just the code.
The frontend technically works but looks wrong, drifts from design intent, or breaks at specific viewport sizes.
Memory Bank Continuity
Give stateless agents a project spine.
Agent work spans multiple sessions, contributors, or tools and context keeps getting rebuilt from scratch.
Sandboxed YOLO Probe
Let the agent run wild somewhere boring.
You want a coding agent to iterate freely, but the command surface, repo contents, or prompt-injection risk makes direct host execution unsafe.
Parallel Agent Worktree Sweep
Run more agents, but keep their blast radiuses separate.
You have multiple independent coding tasks and the human bottleneck is orchestration, not implementation speed.
Agent Merge Queue Review
Separate useful agent output from merge-shaped confetti.
Agent throughput created more branches than a human can safely review from memory.
Completion Promise Loop
Make the agent promise completion against explicit evidence, not vibes and a cheerful summary.
A coding agent keeps declaring work done after the easy part, or a feature has multiple acceptance criteria that must be proven before handoff.
Release Notes Truth Sweep
Turn internal change lists into release notes that match the product users can touch.
Release notes are being drafted from PR titles, tickets, or internal updates and may overstate what users actually get.
Prompt Regression Suite
Treat prompts like code: keep a small regression set before tuning the clever wording.
A prompt, system instruction, tool policy, or model setting change could silently break workflows that used to work.
Meeting Transcript Action Miner
Turn messy meeting records into owned next actions without inventing commitments.
A meeting transcript, chat log, or call note contains promises and decisions that could disappear into the transcript graveyard.
Feature Flag Retirement Sweep
Retire stale flags before yesterday's rollout logic becomes tomorrow's haunted house.
Feature flags, experiments, or kill switches have outlived their rollout and are making code paths harder to reason about.
Schema Migration Dry Run
Make the database change prove itself somewhere safer than the real thing.
A schema change, backfill, cleanup script, or bulk edit could corrupt data, lock tables, or surprise users if run directly in production.
Longform to Asset Pack
Turn one dense source into a small, traceable distribution kit instead of posting chopped-up leftovers.
A valuable longform source should become multiple useful public artifacts without flattening nuance or leaking private context.
Incident Timeline Reconstruction
Turn incident noise into a usable timeline before the story hardens into folklore.
An outage, degraded workflow, or operational scare has evidence scattered across alerts, logs, chats, dashboards, and human memory.
Customer Proof Packet
Package customer evidence into useful proof without laundering private details into public copy.
A promising customer story, testimonial, support win, or product result could become marketing proof but contains private context or unsupported claims.
RAG Grounding Audit
Test the retrieval chain before polished nonsense reaches users.
A chatbot, agent, search page, or knowledge workflow answers from docs and could sound confident while grounded in the wrong source.
Side-Effect Dry Run Gate
Make the agent show its hands before it touches reality.
An automation can send, post, delete, buy, invite, merge, schedule, unsubscribe, or change account, product, customer, or repo state.
Vendor Security Packet Builder
Turn scattered security facts into a reusable response packet without improvising compliance poetry.
A prospect, partner, or platform asks for security, privacy, compliance, or infrastructure details and the answer could overpromise or expose sensitive internals.
Model Output Drift Watch
Catch quiet AI behavior drift before users become the monitoring system. They charge more.
A production AI workflow can change behavior because the model, prompt, tool results, retrieval index, or upstream data shifted under it.
Webhook Replay Harness
Turn unreliable webhook mysteries into replayable fixtures before touching live traffic.
An integration bug depends on third-party webhook payloads, retry timing, signatures, or idempotency behavior that is hard to reproduce from normal tests.
Stakeholder Briefing Packet
Compress scattered context into a source-backed brief so the human shows up oriented, not archaeology-adjacent.
A meeting or update matters enough that walking in cold would waste the room or create avoidable relationship debt.
Bug Report to Repro Kit
Turn blurry bug reports into a small, executable evidence packet before anyone starts spelunking.
A bug report has real signal but is missing the exact steps, inputs, logs, screenshots, account state, or acceptance test an engineer needs to act without guessing.
Recurring Question to FAQ
Convert repeated confusion into one reusable answer instead of playing inbox whack-a-mole with nicer fonts.
People keep asking the same product, pricing, setup, troubleshooting, or workflow question and the answer is scattered across notes, docs, Slack, or someone's head.
Permission Scope Minimizer
Shrink the blast radius before the integration becomes load-bearing and mysteriously sacred.
A tool, plugin, OAuth app, API key, or agent workflow asks for broad access and nobody has proven which permissions are actually needed.
Notebook to Repeatable Pipeline
Turn fragile analyst archaeology into a repeatable pipeline with receipts.
A one-off notebook, spreadsheet, SQL query, or script has become a recurring report or decision workflow.
CRM Hygiene Loop
A bounded loop for when contacts, stages, and follow-ups decay after busy weeks.
Contacts, stages, and follow-ups decay after busy weeks.
Design Token Drift Loop
Find hard-coded style values that should be tokens, replace one cluster per round, and stop only when token usage, component examples, and screenshots agree.
Product screens still ship hard-coded colors, spacing, type, or radii that ignore the token file, so the design system and the live UI disagree.
Support Macro Learning Loop
A bounded loop for when repeated support answers should become product fixes, macros, or docs.
Repeated support answers should become product fixes, macros, or docs.
Agent Cost Budget Loop
A bounded loop for when agent runs consume meaningful tokens without clear ROI.
Agent runs consume meaningful tokens without clear ROI.
SEO Monthly Rank Experiment
Scheduled rank-KPI SEO loop: one experiment per cycle, measure next tick, undo losers.
You want search (and LLM/GEO) visibility to climb over months without hiring an agency, and you can give an agent read access to rank data plus write access to a small content surface.
Paid Ads Copy Variant Loop
Ads loop: generate copy variants, spend enough for signal, kill losers, scale winners.
Paid search or social ads are running and performance is flat, and you need a disciplined cut-losers / double-winners loop without letting the agent freely blow budget.
Product Feedback to Measure Loop
Turn feedback + analytics into one measured product change per cycle; keep bugs separate.
Customer feedback and product telemetry are piling up, but shipping is driven by loudest request instead of a measured build-learn cycle.
No loops match that filter. Either you found a gap, or search is being dramatic.