AI Engineering Operating Library

A new way to work with AI across every technical area

Logic Basics gives teams 74 structured skills for Codex, Claude Code, Cursor, Copilot, ChatGPT, and similar agents, covering delivery, architecture, security, data, research, scale readiness, onboarding, and diagnosis. The basic pack is pay what you want, and the premium pack is available through Gumroad for a simple one-time purchase.

Portable .skills folder pattern. Works across the tools your team already uses.

.skills/master-skills
1# Logic Basics
2# AI engineering operating library
3
4copy .skills/scalability-failure-profiler.md
5copy .skills/gap-discovery-facilitator.md
6
7// 74 ready-to-use skills available
8// 17 workflow chains available
9// Review evidence required: YES
10// Codex, Claude, Cursor, Copilot: READY

74

Total skills

40

Premium skills

17

Workflow chains

9

Categories

It is not a Prompt Pack

Logic Basics is a new way to work with AI: reusable operating procedures for each area of engineering and research, with validation evidence built in.

74 Ready-to-Use Skills

Structured Markdown skills for frontend, backend, database, data science, learning, research, fullstack, utility, and project management work.

Base and Premium Tiers

Use base skills for focused repeatable tasks and premium skills for higher-risk architecture, governance, scale, and diagnosis.

Diagnostic Growth Workflows

Answer questions like what is missing, what am I not seeing, what breaks at 100k users, and how should juniors start.

Didactic Usage Guide

Clear instructions for copying skills into a project-level .skills folder and prompting Codex, Claude, Cursor, Copilot, or ChatGPT.

Validation and Catalog

A generated skills-catalog.json and validator script keep metadata, README coverage, version tracks, and skill structure aligned.

Recommended Chains

Pre-build diagnosis, scale readiness, junior ramp-up, feature delivery, security review, and data quality workflows are ready to run.

Not Prompts. Area Operating Systems.

Each area gets focused skills, recommended chains, and review standards. The point is not to ask AI harder; it is to make AI-assisted work easier to trust.

Area guide

Front-end

Pattern inventory, accessible components, forms, responsive layouts, design tokens, didactic guidance, and Storybook coverage.

ui-pattern-inventory-mapper -> design-system-didactic-centralizer -> storybook-automator

Area guide

Back-end

REST contracts, webhooks, event-driven flows, authentication, consistency, and observability.

webhook-contract-guardian -> event-driven-orchestrator -> observability-stack-composer

Area guide

Database

Migrations, query tuning, ORM modeling, rollback strategy, CDC, and data movement.

migration-safety-checker -> sql-query-tuning-doctor -> database-version-rollback

Area guide

Data Science

EDA, experiment evaluation, KPI stories, feature stores, forecasting, and governed ML work.

exploratory-data-analysis-guide -> ab-test-evaluator -> ml-experiment-tracking-governor

Area guide

Learning

Programming concepts, unfamiliar languages, didactic code, exercises, rubrics, and lesson paths.

language-ramp-up-guide -> programming-concept-explainer -> exercise-lab-builder

Area guide

Research

Literature evidence, protocols, notebooks, scientific claims, grants, physics, and bioinformatics.

literature-evidence-mapper -> scientific-claim-auditor -> grant-manuscript-reviewer

Area guide

Project Management

Gap discovery, hidden requirements, readiness, capacity, dependencies, roadmaps, and onboarding.

gap-discovery-facilitator -> hidden-requirement-radar -> scope-readiness-checker

Area guide

Utility

Repository intake, architecture concerns, testing, commits, docs, environment config, CI/CD, Docker, observability, and scale readiness.

repository-intake-cartographer -> architecture-concern-radar -> cicd-security-hardener

Area guide

Fullstack

Vertical slices across API, data, UI, state, validation, tests, security, and release notes.

vertical-slice-feature-craft -> unit-test-coverage-forcer -> cicd-security-hardener

Built for Teams Already Using AI

The strongest buyers are not asking whether AI can write code. They are asking how to make AI-assisted work consistent, reviewable, and safe to scale.

Software Agencies

Standardize AI-assisted delivery across client projects without forcing every developer to invent their own prompting style.

  • Reusable kickoff and delivery workflows
  • Consistent review evidence
  • Safer junior contribution paths

Startup CTOs

Run senior-level diagnosis before scaling, shipping risky features, or committing to unclear technical plans.

  • Gap and hidden requirement discovery
  • 100k-user failure profiling
  • Architecture and operational trade-offs

Engineering Managers

Give teams shared AI operating standards for implementation, validation, onboarding, observability, and governance.

  • Project-level .skills installation
  • Repeatable workflow chains
  • Less reviewer guesswork

Research Teams

Bring evidence mapping, claim review, reproducibility checks, and manuscript revision workflows into AI-assisted science work.

  • Literature and evidence mapping
  • Scientific claim auditing
  • Notebook and model reproducibility

How It Works

A small project-level skill pack turns vague agent requests into repeatable team behavior.

01

Choose the Workflow

Start from a real engineering question: what is missing, what breaks first, how do we test this, or how should a junior begin?

02

Install Into .skills

Copy selected Markdown skills into a project-level .skills folder so every agent and teammate can reuse the same operating procedure.

03

Require Evidence

Ask the agent to follow the skill workflow and return decisions, changed files, validation evidence, risks, and next steps.

The Difference Is Specificity

Weak prompts create generic output. Skill-guided prompts force assumptions, trade-offs, validation evidence, risks, and a reviewable result.

Weak prompt

Can this app handle 100k users?

Skill-guided prompt

Apply scalability-failure-profiler. State traffic assumptions, rank bottlenecks, list missing metrics, and propose mitigation order.

Workflow chain

Pre-build diagnosis: gap-discovery-facilitator -> hidden-requirement-radar -> scope-readiness-checker.

Browse All Workflows

Use these chains when one skill is too narrow. Each card includes the recommended sequence and a copyable starter prompt.

New fullstack feature

When: Use when a feature touches API, persistence, UI, tests, and release safety.

vertical-slice-feature-craft -> unit-test-coverage-forcer -> storybook-automator -> cicd-security-hardener

Use this workflow: vertical-slice-feature-craft -> unit-test-coverage-forcer -> storybook-automator -> cicd-security-hardener to run the new fullstack feature workflow for [feature]. Return changed files, tests run, validation evidence, risks, and changelog notes.

Incremental modernization

When: Use when improving old code without changing behavior or taking a risky rewrite path.

legacy-code-modernizer -> docstring-artisan -> conventional-commit-helper

Use this workflow: legacy-code-modernizer -> docstring-artisan -> conventional-commit-helper to run the incremental modernization workflow for [module]. Preserve behavior, modernize in small steps, add useful docstrings where they reduce maintenance risk, then prepare a conventional commit and changelog summary.

Data performance

When: Use when slow data access may require schema, query, ORM, or caching changes.

prisma-typeorm-generator -> sql-query-tuning-doctor -> cache-strategy-selector

Use this workflow: prisma-typeorm-generator -> sql-query-tuning-doctor -> cache-strategy-selector to run the data performance workflow for [query or endpoint]. Review ORM model/query shape, diagnose SQL performance, then recommend caching only where it has clear invalidation and consistency rules.

Secure external integration

When: Use when connecting to third-party systems, webhooks, queues, or async delivery.

webhook-contract-guardian -> event-driven-orchestrator -> observability-stack-composer

Use this workflow: webhook-contract-guardian -> event-driven-orchestrator -> observability-stack-composer to run the secure external integration workflow for [integration]. Validate contracts, idempotency, retries, event orchestration, failure handling, observability, and operational alerts.

Delivery planning

When: Use before committing teams to dates, dependencies, sprint goals, or roadmap order.

backlog-refinement-facilitator -> sprint-capacity-planner -> cross-team-dependency-mapper -> roadmap-priority-orchestrator

Use this workflow: backlog-refinement-facilitator -> sprint-capacity-planner -> cross-team-dependency-mapper -> roadmap-priority-orchestrator to run the delivery planning workflow for [initiative]. Refine backlog items, estimate capacity, map cross-team dependencies, and recommend roadmap priority with trade-offs and risks.

Data science delivery

When: Use when analysis must become a decision, metric, dashboard, or governed experiment.

exploratory-data-analysis-guide -> ab-test-evaluator -> kpi-dashboard-storyliner -> ml-experiment-tracking-governor

Use this workflow: exploratory-data-analysis-guide -> ab-test-evaluator -> kpi-dashboard-storyliner -> ml-experiment-tracking-governor to run the data science delivery workflow for [dataset or experiment]. Produce EDA findings, experiment validity checks, KPI dashboard story, and tracking/governance recommendations.

Scientific research review

When: Use when papers, grants, or manuscripts need evidence mapping, claim calibration, reviewer objections, and revision strategy.

literature-evidence-mapper -> scientific-claim-auditor -> grant-manuscript-reviewer

Use this workflow: literature-evidence-mapper -> scientific-claim-auditor -> grant-manuscript-reviewer to run the scientific research review workflow for [paper, grant, or research question]. Map the evidence, audit claim strength and causal language, identify missing controls or citations, then recommend manuscript or grant revisions with validation evidence.

Computational science reproducibility

When: Use when research data, notebooks, simulations, or calculations need provenance, rerun evidence, unit checks, and sanity validation.

research-data-cleanroom -> computational-notebook-reproducer -> physics-model-sanity-checker

Use this workflow: research-data-cleanroom -> computational-notebook-reproducer -> physics-model-sanity-checker to run the computational science reproducibility workflow for [dataset, notebook, or simulation]. Preserve raw data, document cleaning, make the notebook rerunnable, check model assumptions or units, and return validation evidence plus residual risks.

Unknown project intake

When: Use when you just received a repository or project and need to understand what it is before planning or changing anything.

repository-intake-cartographer -> architecture-concern-radar -> technical-due-diligence-assessor -> project-kickoff-charter-builder

Use this workflow: repository-intake-cartographer -> architecture-concern-radar -> technical-due-diligence-assessor -> project-kickoff-charter-builder to run the unknown project intake workflow for [repository or project]. Scan files and docs first, map the stack, architecture, entrypoints, scripts, dependencies, risks, unknowns, and delivery concerns, then produce an onboarding brief, evidence-backed concern matrix, kickoff charter, and next investigation steps.

Didactic design system centralization

When: Use when front-end patterns are scattered and the team needs a centralized design system that contributors can learn and follow.

ui-pattern-inventory-mapper -> design-token-governor -> design-system-didactic-centralizer -> storybook-automator

Use this workflow: ui-pattern-inventory-mapper -> design-token-governor -> design-system-didactic-centralizer -> storybook-automator to centralize the front-end design system for [project]. Inventory current UI patterns, normalize tokens and component rules, create didactic usage guidance for contributors, define migration steps, and add Storybook/documentation coverage with validation evidence.

New language ramp-up

When: Use when a student or developer needs to work in a programming language, framework, or repository they do not know yet.

repository-intake-cartographer -> language-ramp-up-guide -> programming-concept-explainer -> docstring-artisan

Use this workflow: repository-intake-cartographer -> language-ramp-up-guide -> programming-concept-explainer -> docstring-artisan to run the new language ramp-up workflow for [language, repository, or task]. Map the project and language conventions, explain unfamiliar syntax and concepts, identify safe first tasks, and add learner-friendly documentation with validation evidence.

Didactic code simplification

When: Use when code needs to be easier for students, juniors, or new language learners to understand without changing behavior.

programming-concept-explainer -> didactic-code-simplifier -> docstring-artisan

Use this workflow: programming-concept-explainer -> didactic-code-simplifier -> docstring-artisan to run the didactic code simplification workflow for [code sample or module]. Explain the concept, simplify the code while preserving behavior, add useful comments or docstrings, and return before/after reasoning plus validation evidence.

Teacher exercise studio

When: Use when a teacher or mentor needs to plan, create, validate, or review programming exercises and assessments.

learning-path-curriculum-designer -> exercise-lab-builder -> assessment-rubric-reviewer -> unit-test-coverage-forcer

Use this workflow: learning-path-curriculum-designer -> exercise-lab-builder -> assessment-rubric-reviewer -> unit-test-coverage-forcer to run the teacher exercise studio workflow for [topic, language, or class]. Plan the learning path, create exercises with starter code, hints, tests, teacher notes, and rubrics, then review assessment fairness and validation evidence.

Project start and evaluation

When: Use when starting a project, evaluating a client request, or deciding whether work is ready to build.

project-kickoff-charter-builder -> scope-readiness-checker -> technical-due-diligence-assessor -> delivery-feasibility-evaluator

Use this workflow: project-kickoff-charter-builder -> scope-readiness-checker -> technical-due-diligence-assessor -> delivery-feasibility-evaluator to run the project start and evaluation workflow for [project]. Build the kickoff charter, assess scope readiness, perform technical due diligence, and evaluate delivery feasibility.

Diagnostic pre-build review

When: Use when someone asks 'what is missing?' or 'what necessity am I not seeing?' before implementation.

gap-discovery-facilitator -> hidden-requirement-radar -> scope-readiness-checker

Use this workflow: gap-discovery-facilitator -> hidden-requirement-radar -> scope-readiness-checker to run the diagnostic pre-build workflow for [idea or plan]. Identify visible gaps, hidden requirements, missing owners, risk areas, validation gaps, rollout needs, and readiness blockers.

Scale readiness review

When: Use when asking what breaks first at 100k users or before a growth event.

scalability-failure-profiler -> cache-strategy-selector -> observability-stack-composer

Use this workflow: scalability-failure-profiler -> cache-strategy-selector -> observability-stack-composer to run the scale readiness workflow for [system]. State traffic assumptions, rank first failure points, decide whether caching is justified, and define missing metrics, dashboards, alerts, and load-test evidence.

Junior ramp-up

When: Use when bringing juniors into a project without exposing them to unclear or high-risk work.

junior-onboarding-path-builder -> backlog-refinement-facilitator -> conventional-commit-helper

Use this workflow: junior-onboarding-path-builder -> backlog-refinement-facilitator -> conventional-commit-helper to run the junior ramp-up workflow for [team or project]. Create a 30/60/90 path, safe starter tasks, review checkpoints, mentorship rhythm, backlog items, and commit conventions.

Make AI Coding Output Easier to Trust

Install a small skill pack into your project, point your agent at the workflow, and require validation evidence before accepting the result.

Basic is pay what you want. Premium is $9.99 one-time. Checkout powered by Gumroad.