opencode-config-template

Deploy and troubleshoot Next.js 16+ applications on AWS Amplify Hosting — build spec (amplify.yml), SSR Lambda env-var injection, CloudFront OAC, Route53 DNS, GitHub Actions deploy triggers, post-deploy verification, and rollback strategy

Design and document APIs — REST conventions, OpenAPI/Swagger spec generation, GraphQL schema patterns, API versioning, pagination, rate limiting, error response formats, and HATEOAS

Implement authentication and authorization patterns — OAuth2/OIDC flows, JWT best practices, session management, RBAC/ABAC, NextAuth/Auth.js, Passport.js, password hashing, and CSRF protection

Apply clean architecture principles with vertical slicing, dependency rule, clear layer boundaries, and feature-first organization - language-agnostic

Write clean, human-readable code with proper naming, small functions, self-documenting patterns, and object calisthenics - language-agnostic

Detect and fix code smells including long methods, large classes, feature envy, primitive obsession, and more with refactoring guidance - language-agnostic

Apply GoF design patterns (Creational, Structural, Behavioral) appropriately without over-engineering - language-agnostic with practical examples

Create and optimize Dockerfiles, multi-stage builds, docker-compose patterns, image size reduction, layer caching, .dockerignore, health checks, and container security scanning with hadolint and docker scout

Identify and fix performance bottlenecks — profiling (cProfile, Chrome DevTools), caching strategies (Redis, memoization, CDN), bundle size analysis, lazy loading, N+1 query detection, and memory leak detection

Scaffold and structure Python backend projects with FastAPI, Django, or Flask — project layout, dependency injection, configuration management, virtual environments, and pyproject.toml standards

Detect and fix React 19 and Next.js 16 runtime anti-patterns — hydration mismatches, StrictMode double-execution, memory leaks from module-scope caches, stale derived state, fail-open RBAC, revalidatePath error swallowing, fragment keys, useMemo dep issues, and more

Audit code and dependencies for security vulnerabilities — OWASP Top 10, dependency scanning (npm audit, pip-audit), secret detection, input validation, XSS/CSRF prevention, security headers, and HTTPS enforcement

Generic test generation framework supporting multiple languages and testing frameworks

Generate low-fidelity, hand-drawn web wireframes and clickable prototypes. Use when the user wants to wireframe, mockup, sketch, or prototype a UI before writing production code. Triggers on: wireframe, mockup, lo-fi prototype, rough layout, Balsamiq-style, hand-drawn UI, clickable prototype, sketch the screens, visualize the flow, show what the app would look like, responsive baseline, breakpoint wireframe. Produces self-contained SPAs with sketchy aesthetics — wired-elements components, graph-paper background, doodle icons. Also used as a baseline source for responsive audits: generate wireframes at specific viewport breakpoints to establish the expected structural layout.

Extract and store reusable patterns, decisions, and insights from coding sessions with confidence scoring, project scoping, and instinct evolution for future reference and skill improvement

Actively build and sharpen a project's domain model. Use when the user wants to pin down domain terminology or a ubiquitous language, record an architectural decision, or when another skill needs to maintain CONTEXT.md and ADRs inline. Changes the model — not merely reads it.

Write concise, compact Git commit messages within strict word and character budgets. Triggers when user says: "compact commit", "short commit message", "concise commit", "commit word limit", "brief commit", "terse commit", "make commit shorter", "git commit best practices". Combines related changes into single commits by semantic grouping, enforces 150-word body and 72-char subject limits, and provides commitlint CI/CD configuration. Use for brief, space-efficient commits — NOT for conventional commit format/structure (use git-semantic-commits-skill for that).

Format and write Git commit messages following the Conventional Commits specification. Triggers when user says: "commit and push", "commit changes", "make a commit", "commit this", "write a commit message", "create a commit", "git commit", "commit message format", "semantic commit". Detects commit type (feat, fix, docs, refactor, etc.), adds scopes, identifies breaking changes, and provides semantic versioning guidance. Use for structuring commit messages properly — NOT for making commits shorter or more compact (use git-compact-commits-skill for that).

A relentless interview to sharpen a plan or design. User-invoked orchestrator that runs grilling-skill WITHOUT doc capture. Use when you want a grilling session but don't need CONTEXT.md or ADR artifacts.

A relentless interview to sharpen a plan or design, which also creates docs (CONTEXT.md glossary and ADRs) as we go. User-invoked orchestrator that pairs grilling-skill with domain-modeling-skill.

Interview the user relentlessly about a plan or design. Use when the user wants to stress-test a plan before building it, resolve a design decision tree, or uses any 'grill' trigger phrases. Asks one question at a time with a recommended answer.

Three.js + Next.js (App Router / React 19) integration guidance with MANDATORY version detection. Covers SSR/Server Component pitfalls, Turbopack vs Webpack GLSL/bundler issues, hydration mismatches, StrictMode WebGL context loss, persistent canvas across routes, R3F (React Three Fiber) + drei ecosystem, WebXR/AR/VR via @react-three/xr, and the companion library decision tree (drei, postprocessing, rapier, csg, leva, maath, meshline, troika-three-text, three-mesh-bvh, gltfjsx, zustand). Use when working with three.js, threejs, react-three-fiber, R3F, WebGL, WebXR, ARButton, VRButton, or @react-three/* packages in a Next.js or React 19 project.

Autonomous code optimization loop — modify → verify (benchmark) → keep/revert against any code metric (test coverage, bundle size, runtime, error count). TDD-flavored: test-pass maps to pass, pass-count maps to score.

Canonical source of the autoresearch iteration protocol — a 5-stage Understand → Hypothesize → Experiment → Evaluate → Log & Iterate loop driven by a mechanical {"pass":bool,"score":N} evaluator, with git-as-memory keep/revert, stuck detection, and prompt-injection boundaries. Cited by all domain and retrofitted skills.

[Requires NVIDIA GPU] Autonomous ML training research loop — modify train.py, run a fixed-time-budget experiment, parse val_bpb, keep/revert via git-as-memory. Adapted from karpathy/autoresearch.

Autonomous literature-review and paper-synthesis loop. Tier 2 (web-only, no Bash) — fetches papers, extracts structured summaries, builds a living research.md with categories-covered tracking. Uses agent-as-evaluator fallback when no mechanical evaluator applies.

Ensure test coverage percentage is displayed in README.md for Next.js and Python projects following industry standards

Find, classify, and remove @deprecated code from a TypeScript/Next.js codebase using dependency-traced analysis. Scans for @deprecated markers, traces the full dependency graph (including transitive chains through other deprecated code), classifies items into tiers by deletion safety, and executes phased removal with typecheck verification at each step. Triggers on: deprecated cleanup, remove deprecated, dead code cleanup, deprecated functions, remove @deprecated, cleanup deprecated code, find deprecated, deprecated removal.

Audit and auto-fix documentation consistency across PLAN files, README.md, AGENTS.md, and deploy scripts — cross-file count sync, PLAN vs reality drift, structural integrity, and orphan reference detection.

Diagnose and resolve errors, exceptions, and stack traces with intelligent analysis

Evaluate code quality, skill effectiveness, and implementation correctness against defined criteria with scoring and improvement suggestions

Create distinctive, production-grade frontend interfaces with high design quality that avoid generic AI aesthetics. Use this skill when building web components, pages, landing pages, dashboards, React components, HTML/CSS layouts, or when styling/beautifying any web UI.

Generic linting workflow for multiple languages with auto-fix and error resolution

Create Mermaid diagrams using the Mermaid MCP server — SVG by default, no CLI install needed

Next.js PR workflow with quality checks (lint/build/test), semantic versioning labels (major/minor/patch), coverage badges, and JIRA integration via pr-creation-workflow framework

Generate comprehensive unit tests for Next.js 16 applications covering App Router, Server Components, Client Components, API routes, and Server Actions with industry best practices

Scan, validate, and audit OpenCode skills for consistency, redundancy, and modularization opportunities

Execute PLAN.md phases with automatic progress tracking. Parses plan, executes tasks sequentially, and auto-invokes plan-updater after each phase completion.

Methodology for auditing and fixing responsive UI defects in Next.js projects using Playwright. Defines 6 detection assertions, 3 fix-confidence tiers, and a closed detect-fix-re-verify loop. Loaded by responsive-audit-subagent. Not auto-triggered — invoked exclusively via permission.skill by the audit subagent.

Framework for creating PRs with automatic framework/language detection, configurable quality checks, multi-platform integration, semantic versioning labels (major/minor/patch), JIRA image attachments, and JIRA/git issue tracking