AI & agents
The current acceleration layer. I use AI to inspect codebases, draft plans, prototype faster, and automate the repetitive parts while keeping product judgment human.
Agentic implementation, refactors, repo navigation, PR prep, and verification loops.
Product thinking, debugging, technical planning, copy exploration, and alternate approaches.
Research, long-context review, and a useful second model when answers need pressure-testing.
Daily AI coding environment for frontend, backend, and small automation work.
Connecting AI workflows to GitHub, Supabase, Postman, browser state, and local project context.
Fast research, competitive scans, and source-backed sanity checks.
Rapid UI generation for exploring product directions before rebuilding the durable parts.
Browser-based app experiments when I want a working sketch quickly.
Prompt-to-app prototyping for rough product tests and disposable concept work.
Quick environment experiments, prototypes, and throwaway proofs of concept.
Inline coding assistance, repo-aware suggestions, and lightweight pair-programming support.
Another agentic IDE to compare planning, edits, and multi-file implementation behavior.
Artifacts you can use.
Practical downloads, starter kits, and notes you can use directly, including markdown artifacts made to drop into your AI agent of choice as project context.
AI Evaluation Measurement Contract
A pre-run contract connecting an AI product claim to populations, metrics, graders, uncertainty, failure severity, and release authority.
Dependency Adoption Receipt
A reviewable receipt for package need, identity, provenance, permissions, supply-chain risk, verification, ownership, and removal.
Skills Transfer Evidence Map
A candidate and recruiter map from target-role outcomes to transferable evidence, context differences, structured prompts, and confidence.
AI Interface State Contract
A product and frontend contract for streaming, tool use, uncertainty, recovery, review, and accessible AI interaction states.
AI Evaluation Dataset Starter
A starter structure for test cases, rubrics, failure labels, dataset slices, human calibration, and AI regression review.
Human Review Escalation Matrix
A decision matrix for when AI can act, when it needs confirmation, and when a qualified human must take over.
AI Product Sprint Checklist
A practical sprint checklist for using AI across discovery, UX, implementation, and verification without skipping product judgment.
Design-to-Code Handoff Checklist
A handoff checklist for turning Figma screens into build-ready components, tokens, states, and responsive requirements.
AI Feature UX Checklist
A UX audit for AI-powered features covering trust, uncertainty, fallback states, controls, and evaluation loops.
Landing Page Conversion Checklist
A conversion-focused pass for landing pages across message, hierarchy, forms, analytics, SEO, and performance.
Personal Site Content Audit Template
A portfolio audit template for sharpening positioning, credibility, proof, content structure, and recruiter-facing signals.
Prompt Library for UI Critique
Reusable prompts for pressure-testing layout, copy, hierarchy, accessibility, interaction states, and implementation risk.
Product Analytics Event Taxonomy
A naming and planning template for defining product events, properties, funnels, activation signals, and instrumentation ownership.
Funnel Audit Worksheet
A worksheet for diagnosing acquisition, activation, conversion, retention, and measurement problems in a product funnel.
Shopify App Onboarding Checklist
A commerce-focused onboarding checklist for helping merchants reach first value inside a Shopify app.
Postman API Review Checklist
A review checklist for API collections, auth, examples, edge cases, docs, testing, and agent-ready API behavior.
MCP Starter Map
A map for connecting AI agents to local code, GitHub, Supabase, Postman, browser context, and deployment workflows.
Recruiter-Facing AI Workflow Deck
A concise slide-style walkthrough of how JP uses AI across research, design, engineering, QA, and delivery.
UI PR Risk Review Checklist
A merge-readiness checklist for product intent, states, accessibility, visual durability, and UI implementation risk.
Front-End State Recipes
Reusable recipes for optimistic actions, loading, empty, error, data-transition, and disabled-control states.
Design System Contribution Pack
A contribution brief, drift diagnosis, escape-hatch rules, and component-docs template for product teams.
Product Spec Agent Template
A pasteable agent-context template for product specs, constraints, states, acceptance criteria, and QA.
Roadmap Prioritization Canvas
A decision canvas for comparing build, buy, integrate, defer, and remove options with the same criteria.
Portfolio Case Study Proof Template
A case-study structure for proving judgment, constraints, tradeoffs, messy-middle artifacts, and outcomes.
Handoff Notes Template
A build-ready handoff format for scope, states, interactions, open questions, analytics, and QA.
Design Tokens Starter JSON
A public token starter with JSON source tokens, generated CSS variables, light/dark modes, and a plain HTML example.
@jpcasa/tokens
A published token package with JSON source tokens, generated CSS variables, and package exports for product UI systems.
Casabianca UI Mono / Display Experiment
An experimental type specimen with generated TTF, WOFF, and WOFF2 files for mono UI labels, numerals, and interface previews.
Vue Component Starter Kit
A Stripe-inspired Vue 3 starter with typed SaaS components, tokens, responsive grids, and a polished dashboard example.
React Dashboard Shell
A polished React dashboard starter with sidebar navigation, metrics, filters, tables, detail panels, and reusable UI states.
Agent-Ready API Spec Template
An OpenAPI and Postman starter template for APIs that AI agents can discover, call, and recover from safely.
Supabase Static CMS Starter
An Astro static CMS starter with Supabase content tables, published-only RLS, explicit Data API grants, and fallback content.
Modern tools, senior judgment.
I like new tools when they compress the path from thought to artifact. The trick is knowing what to automate, what to verify, and where taste still has to do the work.