PAI-zong-NB-v5.8

Route user requests to the right Codex role, output mode, and minimal skill set. Use for every task before deciding whether the answer is simple consultation, architecture, project assistance, coding, debugging, review, documentation, or rule maintenance.

Inspect and summarize project context before coding. Use when Codex needs to modify a repository, understand structure, follow existing conventions, locate scripts, read AGENTS.md, README, package manifests, or project documentation.

Design practical user interfaces, flows, information architecture, layout systems, component states, and design tokens. Use for page design, UX redesign, dashboards, tools, mobile/web UI, interaction design, and before frontend implementation when visual or workflow decisions matter.

Build and modify frontend pages, components, state logic, styling, forms, client-side interactions, and UI behavior in React, Vue, Svelte, vanilla HTML/CSS/JS, or similar frameworks. Use for frontend implementation and frontend bug fixes.

Implement backend services, business logic, API handlers, jobs, integrations, file processing, authentication flows, and server-side fixes. Use for backend coding and debugging in Node, Python, Go, Java, Rust, PHP, or similar stacks.

Design REST, GraphQL, RPC, webhook, and internal service APIs. Use for routes, status codes, request/response schemas, error models, pagination, filtering, versioning, idempotency, and OpenAPI or contract documentation.

Design and modify database schemas, migrations, SQL queries, indexes, transactions, constraints, and query performance. Use for PostgreSQL, MySQL, SQLite, SQL Server, Prisma, ORM changes, migration planning, and data consistency issues.

Analyze and implement security-sensitive code involving authentication, authorization, tokens, sessions, uploads, payment, encryption, secrets, webhooks, sensitive data, dependency risk, injection, XSS, CSRF, SSRF, SQL injection, and audit logging.

Create, repair, and evaluate tests for unit, integration, end-to-end, contract, regression, CI, and bug-fix workflows. Use when implementing code with meaningful risk, fixing bugs, reviewing test gaps, or investigating failing test suites.

Work on Docker, deployment, CI/CD, environment variables, server configuration, build pipelines, health checks, logs, monitoring, rollbacks, infrastructure scripts, and production operations.

Check and improve accessibility for web and app UI: semantic HTML, keyboard navigation, focus management, accessible names, ARIA, forms, dialogs, error messaging, color contrast, motion preferences, and WCAG 2.2 AA-oriented implementation.

Diagnose and improve frontend, backend, database, build, memory, CPU, and network performance. Use for slow pages, slow APIs, slow SQL, large bundles, rendering delays, high latency, resource spikes, or performance regressions.

Internationalize and localize user interfaces, APIs, documents, and formatting. Use for multilingual apps, locale-sensitive dates, numbers, currency, pluralization, translation keys, lang/dir attributes, RTL layouts, and avoiding hardcoded text.

Create and maintain README files, API docs, usage docs, comments, AGENTS.md, .ai_project.md, prompt rules, skill files, and project conventions. Use for documentation, workflow rules, prompt engineering, knowledge distillation into skills, and Codex behavior refinement.

Review code changes for bugs, regressions, security issues, missing tests, API contract breaks, data risks, performance risks, maintainability problems, and release blockers. Use when the user asks for review, audit, risk check, PR review, or to inspect a diff.

Improve visual quality, spacing, hierarchy, typography, color, motion, icons, empty states, and perceived product quality for desktop and web UI. Use when the user asks for UI beautification, visual polish, making an interface look professional, modernizing screens, or improving design quality without changing core functionality.

Build and maintain API client layers for desktop and web apps, including request wrappers, typed clients, authentication headers, retries, cancellation, streaming, pagination, error normalization, rate limits, and API-driven UI state.

Build and modify API-driven desktop applications using Electron, Tauri, Qt, .NET, native shells, or similar stacks. Use for desktop UI, native capabilities, local storage, auto-update, packaging, permissions, IPC, filesystem access, and desktop-specific UX.

Build and improve API-driven web applications, dashboards, admin panels, SaaS tools, and browser-based product UI. Use for routing, app shell, auth-gated pages, data views, tables, filters, forms, responsive layouts, and production web UX.

Set up and use 1Password CLI (op). Use when installing the CLI, enabling desktop app integration, signing in, and reading/injecting secrets for commands.

Design and implement client-side data fetching and server state for API-driven desktop and web apps: caching, invalidation, optimistic updates, polling, subscriptions, pagination, stale data, offline behavior, and request race handling.

Implement authentication and authorization integration for API-driven desktop and web apps: login, OAuth/OIDC, API keys, token refresh, session expiry, RBAC/ABAC, protected routes, desktop credential storage, and auth error UX.

Package, release, and update desktop and web applications: Electron/Tauri builds, installers, auto-update, code signing, web builds, environment config, versioning, release notes, rollback, and distribution checks. Use with ship-pro for release readiness.

PostgreSQL-specific professional development and optimization. Use for Postgres schema design, migrations, indexes, EXPLAIN, JSONB, full-text search, constraints, transactions, locks, isolation, RLS, partitioning, generated columns, extensions, and query tuning.

Disciplined debugging and diagnosis workflow for hard bugs, regressions, flaky behavior, performance issues, and unknown failures. Adapted from Matt Pocock-style diagnosis loops for Codex v4.1. Use when root cause is unclear or the bug is not fixed by a direct local change.

Test-driven development workflow using red-green-refactor for features, bug fixes, and risky behavior changes. Adapted from Matt Pocock-style TDD for Codex v4.1. Use when tests can clarify requirements, prevent regressions, or guide implementation in small vertical slices.

Build fully-integrated 3-statement models (IS, BS, CF) in Excel with working capital schedules, D&A roll-forwards, debt schedule, and the plugs that make cash and retained earnings tie. Pairs with excel-author.

Review and improve codebase architecture, module boundaries, coupling, domain language, data flow, and long-term maintainability. Adapted from Matt Pocock-style architecture improvement workflows for Codex v4.1. Use for architecture review, refactor planning, complex feature placement, or when code is becoming hard to change.

Professional web application testing with real browser workflows, Playwright-style checks, accessibility smoke tests, responsive screenshots, API mocks, regression coverage, and QA matrices. Use for browser validation, E2E tests, UI regression checks, and release-readiness testing.

Translate Figma designs into production-ready frontend code and design-system rules. Inspired by OpenAI Figma implementation skills and adapted for Codex v4.1. Use when implementing a Figma design, extracting design tokens, matching visual fidelity, or converting design specs to components.

Professional security review and threat modeling for repositories, APIs, desktop apps, web apps, authentication, authorization, secrets, dependency risk, and data flows. Inspired by OpenAI security-best-practices and security-threat-model skills, adapted for Codex v4.1.

Maintain long-lived project context for software projects. Use when creating or modifying project code, adding features, changing architecture, adding files, changing APIs, database schema, deployment, tests, or documentation. Ensures .ai_project.md and docs/ are read before work and updated after meaningful changes.

Manage project evolution over time: feature lifecycle, ADR decisions, roadmap, technical debt, known issues, refactor candidates, release notes, upgrade history, and durable development process records. Use when adding features, planning upgrades, changing architecture, stabilizing a project, or keeping long-term project knowledge current.

Distill books, articles, external repositories, expert workflows, team conventions, and repeated project experience into concise executable Codex skills. Inspired by kangarooking/cangjie-skill and adapted for Codex v4.3. Use when converting knowledge into SKILL.md files, improving skill packs, or extracting reusable methods from references.

High-taste UI judgment for avoiding generic AI-looking interfaces and improving product feel. Inspired by Leonxlnx/taste-skill and adapted for Codex v4.3. Use for UI beauty, frontend aesthetics, visual direction, layout taste, avoiding generic slop, and reviewing whether an interface feels polished or amateur.

Professional UI/UX design review for web apps, desktop apps, dashboards, landing pages, and product screens. Inspired by gstack design-review/design-consultation workflows and adapted for Codex v4.3. Use when reviewing screenshots, implemented UI, Figma designs, visual regressions, or proposed layouts.

Release-quality QA workflow for web and desktop software. Inspired by gstack QA workflows and adapted for Codex v4.3. Use before shipping, after feature completion, for regression sweeps, smoke testing, manual QA plans, automated QA, and acceptance validation.

Shipping and release readiness workflow for software changes. Inspired by gstack ship workflows and adapted for Codex v4.3. Use when preparing a feature, desktop app, web app, API change, or release for delivery, including final checks, docs, changelog, versioning, rollback, and deployment readiness.

Govern a growing Codex skill pack: evaluate external skills, prevent duplicate or conflicting skills, version skill changes, track sources, define skill tiers, retire weak skills, and maintain trigger quality. Use when adding, reviewing, merging, pruning, distributing, or upgrading skills.

Search, evaluate, and select external agent skills from skills.sh, GitHub repositories, curated lists, and local skill packs. Use when the user asks whether external skills are useful, wants to find professional skills, compare skill repositories, or improve the current skill pack with outside sources.