By vTRKA
Supervibe plugin: specialist agents, trigger-safe workflow routing, worktree-ready autonomous loops, design intelligence, code graph, project memory, confidence gates, and scaffolding for 25 stacks.
npx claudepluginhub vtrka/supervibe --plugin supervibeSync project-level host artifacts to upstream plugin changes or user-requested project-fit changes for agents, rules, and skills. Diff-driven, user-gated.
Health-check across project's agents, rules, memory entries, indexes, design-intelligence evidence, and host instruction routing — surfaces stale artifacts and inconsistencies.
Use WHEN starting brainstorm, clarifying a feature, or continuing after an approved brainstorm TO produce an approved spec and hand off with "Next: /supervibe-plan" instead of stopping. Triggers: 'brainstorm', 'next plan', 'брейншторм', 'я сделал брейншторм', 'план'.
Use WHEN checking Codex, Cursor, Gemini, OpenCode, Copilot, Claude Code, IDE registration, plugin manifests, local CLI availability, or host compatibility TO produce a read-only multi-host doctor report with exact repair actions.
Use WHEN an approved implementation plan has passed review and atomic tasks are ready TO execute the plan with explicit readiness and completion gates. Triggers: 'execute plan', 'сделай по плану', 'review passed', 'atomic tasks ready', 'epic ready', '/supervibe-execute-plan'.
Use WHEN cleaning stale completed epics, archived work-item graphs, superseded memory, low-confidence old learnings, stale receipt archives, runtime logs, or .supervibe clutter TO run the reversible Supervibe garbage-collection preview or apply flow.
Bootstrap a host-aware Supervibe scaffold from a stack-pack matched to the detected project stack. Invokes the supervibe:genesis skill end-to-end.
Use WHEN running an autonomous loop, epic, atomic work queue, goal-until-complete session, or separate worktree TO perform provider-safe policy preflight, goal-bounded execution, status, resume, stop, side-effect ledger, and cleanup gates. Triggers: 'loop', 'epic', 'worktree', 'goal-complete', 'эпик'.
Use AFTER approved spec OR WHEN planning is ready TO write a phased plan, require review loop, then atomic task split and epic handoff before execution. Triggers: 'plan', 'review plan', 'atomize', 'сделал план', 'план', 'ревью', 'эпик'.
Create and iterate presentation decks with Supervibe designers: brief -> storyboard -> HTML slide preview -> feedback -> approved PPTX -> Google Drive handoff.
Manage local preview servers for HTML/CSS/JS mockups (start / list / kill). Use after generating mockup files to view them at http://localhost:PORT with auto-reload.
Score an artifact against its confidence rubric with explicit gap analysis and remediation actions. Primary scoring command; use --record when telemetry should be updated. Supports auto-detect, batch, dry-run, and override flow. Triggers: 'оцени', 'score', 'evaluate confidence', 'gap analysis', '/supervibe-score'.
Use WHEN the user asks for security audit, vulnerability scan, safe-to-ship decision, or remediation loop TO run a read-only multi-agent security audit, rank findings, then optionally plan, execute, and re-audit until the security gate reaches 10/10. Triggers: 'security audit', 'vulnerability scan', 'safe to ship', 'секьюрити аудит', 'проверка уязвимостей'.
Use WHEN checking Supervibe health, work-item watch state, delegated inbox, or tracker visibility, dashboard output, external integration readiness, saved views, structured queries, interactive status palette, or local work reports or eval reports TO show a read-only project status report.
Strengthen underperforming agents/skills using telemetry from .supervibe/memory/agent-invocations.jsonl with quantitative metric thresholds. Detects: avg-confidence <8.5 over last 10, override-rate ≥40% delta over 50/50 split, repeated stale-context (≥2/10 invocations). Triggers: 'strengthen agent', 'усиль агента', '/supervibe-strengthen'.
Use WHEN the user wants to see or manage epics, work items, phases, loop state, context packs, blockers, GC previews, or local task actions visually TO launch the local Supervibe UI control plane.
Update the Supervibe plugin: git pull + required ONNX model setup + npm ci + registry rebuild + install doctor. Idempotent. Triggers: 'update plugin', 'обнови плагин', 'supervibe upgrade', '/supervibe-update'.
Use WHEN the user asks what to do next, has a brainstorm or plan ready, asks for review, atomic task split, epic creation, autonomous run, security audit, router/network diagnostics, design/UI/deck requests, worktree isolation, status, stop/resume, or diagnose/why-trigger TO route to the next safe Supervibe workflow command without mutating first. Triggers: 'what next', 'brainstorm', 'plan', 'review', 'security audit', 'брейншторм', 'план', 'ревью', 'эпик', 'уязвимости'.
Use WHEN reviewing changes that affect layer boundaries, dependency direction, or coupling to assess architectural soundness READ-ONLY. Triggers: 'отревьюй архитектуру', 'оцени архитектурное решение', 'проверь дизайн системы', 'архитектурное ревью'.
Use BEFORE designing or modifying authentication/authorization (login, sessions, tokens, MFA, SSO) to choose protocols and prevent common auth flaws. Triggers: 'дизайн авторизации', 'auth flow', 'OAuth схема', 'настрой логин', 'JWT/сессии'.
Use BEFORE merging any change to systematically review code across 8 dimensions with severity-ranked findings. Triggers: 'проверь код', 'код-ревью', 'отревьюй PR', 'review этот код'.
Use AS LAST gate before claiming any work done to verify all evidence present and confidence ≥9 across applicable rubrics. Triggers: 'override review', 'обоснуй gate', 'финальная проверка', 'quality gate'.
Use WHEN improving code structure WITHOUT changing behavior to apply preserve-behavior refactoring with caller-verification via grep and a green-test baseline. Triggers: 'отрефактори', 'переименуй', 'extract method', 'вынеси в функцию', 'упрости код'.
Use BEFORE making changes in unfamiliar code area to map existing structure, patterns, and risks via READ-ONLY exploration backed by supervibe:code-search semantic queries. Triggers: 'разберись в репо', 'research code', 'как устроен этот код', 'изучи проект'.
Use WHEN encountering any bug, test failure, or unexpected behavior to find root cause via hypothesis-evidence-isolation method, never symptom suppression. Triggers: 'почему ломается', 'найди причину', 'дебаг', 'почему не работает', 'тест падает'.
Use BEFORE merging changes touching auth/secrets/data-handling to audit OWASP Top 10 risks, secrets exposure, permissions, and attack surface; also use as the lead auditor for /supervibe-security-audit remediation loops. Triggers: 'аудит безопасности', 'проверь на уязвимости', 'security review', 'проверь секреты', 'safe to ship', 'security audit loop'.
Use BEFORE shipping any UI to verify WCAG AA compliance, keyboard navigation, screen reader support, contrast measurement, motion sensitivity, and ARIA correctness. Triggers: 'проверь доступность', 'a11y review', 'accessibility audit', 'WCAG проверка', 'screen reader test'.
Use WHEN writing or reviewing UI copy (labels, body, CTAs, errors, microcopy) to ensure voice consistency, clarity, and localization-readiness. Triggers: 'напиши тексты', 'отредактируй копи', 'пройдись по текстам', 'CTA подбери'.
Use WHEN starting any new product or major visual direction shift to define brand language, mood, palette intent, typographic intent, motion intent, and emotional anchors. Triggers: 'нужен бренд', 'разработай бренд', 'визуальное направление', 'redesign', 'rebrand', 'фирстиль', 'mood-board', 'дизайн-направление'.
Use WHEN designing, reviewing, or implementing dashboards, charts, maps, analytical UX, metric cards, data storytelling, data-density choices, accessibility fallbacks, and chart library handoffs.
Use WHEN importing, auditing, reconciling, or strengthening design-intelligence datasets, manifests, source variants, CSV domains, reference cards, and design retrieval evidence.
Use WHEN defining, reviewing, extending, or auditing the project design-system contract: tokens, component baseline, motion, accessibility platform rules, component-library bridge, token drift, and prototype-to-production governance.
Use WHEN designing UI for an Electron desktop application — main window, settings, modals, tray dropdowns, multi-window experiences — to produce platform-faithful mockups that respect macOS / Windows / Linux HIG, native title-bar conventions, and keyboard accelerator etiquette. Triggers: 'design Electron app', 'дизайн десктоп-приложения', 'electron UI', 'tray icon design', 'settings window', 'native title bar', 'desktop HIG'.
Use WHEN designing UI for a browser extension (Chrome MV3, Edge, Brave, Firefox WebExtensions) — popup, side panel, options, new-tab override — to produce surface-aware mockups that respect host-browser etiquette, CSP constraints, and platform conventions. Triggers: 'design extension popup', 'дизайн расширения', 'popup для расширения', 'side panel design', 'options page', 'new tab override', 'chrome extension UI', 'MV3 popup'.
Use WHEN designing UI for a native mobile app — iOS (SwiftUI / UIKit), Android (Jetpack Compose / Views), React Native, or Flutter — to produce platform-faithful mockups that respect iOS HIG and Android Material 3, safe-area constraints, one-handed reach, gesture etiquette, and battery budgets. Triggers: 'design mobile app', 'дизайн мобильного приложения', 'iOS app design', 'Android app design', 'React Native UI', 'Flutter UI', 'mobile-native', 'navigation mobile', 'bottom sheet design'.
Use WHEN a storyboard or presentation brief must become a browser-reviewed slide preview, revised through feedback, exported as PPTX, and prepared for Google Drive handoff.
Use WHEN a user needs a persuasive, educational, investor, sales, product, or internal presentation to define audience outcome, narrative spine, slide architecture, visual references, and design-system alignment.
Use WHEN materializing design as 1:1 HTML/CSS prototype in .supervibe/artifacts/prototypes/ for brandbook approval and 1:1 production transfer. Triggers: 'построй прототип', 'свёрстай мокап', 'нужен HTML-прототип', 'кликабельный макет'.
Use WHEN designing UI for a Tauri 2 desktop application — main window, secondary windows, tray, system dialogs — to produce mockups that work identically on WKWebView (macOS), WebView2 (Windows), and WebKitGTK (Linux) without Chromium-only assumptions. Triggers: 'design Tauri app', 'дизайн tauri-приложения', 'tauri UI', 'desktop app lightweight', 'system tray tauri', 'webview2 compatibility', 'WKWebView design'.
Use BEFORE marking any UI implementation done to review across 8 dimensions (hierarchy/spacing/alignment/states/keyboard/responsive/copy/DS-consistency). Triggers: 'отполируй UI', 'review интерфейса', 'проверь дизайн готов ли'.
Use WHEN designing screens or flows to produce screen specs with information architecture, component inventory, states matrix (loading/empty/error/success/partial), interaction notes, design tokens. Triggers: 'спроектируй экран', 'дизайн флоу', 'нужны экраны', 'UI спецификация', 'дизайн интерфейса'.
Use WHEN deciding which Supervibe phase to invoke based on weighted context (system-reminders, effectiveness, confidence-log, user message, stack-fingerprint) — never auto-executes state changes. Triggers: 'распредели задачу', 'кого позвать', 'выбери агента', 'какой агент подходит'.
Use WHEN auditing project memory hygiene OR after sync-rules to deduplicate, retire stale entries, normalize tags, regenerate index, and audit cross-link integrity. Triggers: 'добавь в память', 'сохрани решение', 'memory entry', 'почисти память', 'аудит памяти'.
Use WHEN adding/modifying/auditing/retiring project rules to maintain the selected host rules folder, detect contradictions, normalize format, and sync across sibling repos. Triggers: 'обнови правило', 'rules audit', 'дисциплина', 'добавь правило', 'проверь правила'.
Use WHEN designing or reviewing multi-agent workflows, tool routing, handoffs, agent state, task graphs, autonomous loops, planner/executor splits, or production agent operating models. Triggers: "agent workflow", "multi-agent", "orchestrator", "handoff", "planner", "autonomous loop", "tool routing".
Use WHEN designing LLM/AI integration into product code (prompts, RAG, vector DB, embeddings, model routing, evaluation harnesses, prompt-injection defenses). Triggers: 'интегрируй LLM', 'AI feature', 'prompt design', 'RAG', 'векторная база'.
Use WHEN reviewing API changes (REST/GraphQL/gRPC) to detect breaking changes, version compatibility, and contract drift. Triggers: 'breaking change?', 'отревьюй API', 'версионирование', 'совместимость API'.
Use BEFORE finalizing API contracts (REST/GraphQL/gRPC) to design versioning, error envelopes, idempotency, pagination, and deprecation strategy. Triggers: 'спроектируй API', 'REST/GraphQL дизайн', 'эндпоинты', 'дизайн контракта'.
Use WHEN needing current 2026 best practices for a stack/library to research authoritative sources, cite, and apply to project context. Triggers: 'актуальные best practices', 'docs research', '2026 практики', 'как сейчас принято'.
Use WHEN researching market visual/UX patterns for product category to inform brand and design without copying. Triggers: 'как у конкурентов', 'скрап конкурента', 'дизайн research', 'посмотри как сделано у'.
Use BEFORE designing or evolving the data model (tables, documents, events) to choose normalization, polymorphism, soft delete, and temporal strategy. Triggers: 'дизайн схемы', 'ER модель', 'нормализация', 'спроектируй таблицы'.
Use WHEN reviewing schema changes, migrations, or query patterns to verify safety, performance, replication impact, and lock duration. Triggers: 'отревьюй миграцию', 'проверь схему', 'индексы', 'safe migration'.
Use WHEN evaluating new or upgrading deps to research latest stable, deprecation status, migration guides from authoritative registry sources. Triggers: 'актуальная версия пакета', 'changelog', 'миграция версии', 'стоит ли обновлять'.
Use WHEN adding or auditing dependencies to ensure license compliance, security (CVE), maintenance signals, and supply chain hygiene. Triggers: 'обнови зависимости', 'CVE', 'lock-file аудит', 'проверь пакеты'.
Use WHEN designing CI/CD, runbooks, SLOs, observability, or incident response to ensure reliability and operability. Triggers: 'настрой деплой', 'CI/CD', 'добавь мониторинг', 'runbook', 'SLO'.
Use WHEN designing HA/replication/cache/queue topology to research current vendor-recommended patterns for project's specific versions. Triggers: 'инфра паттерны', 'topology research', 'как Redis рекомендует', 'vendor docs'.
Use WHEN designing infrastructure topology requiring HA, replication, sharding, queueing, or caching to choose patterns matching scale and reliability requirements. Triggers: 'спроектируй инфру', 'topology', 'redis cluster vs sentinel', 'HA схема'.
Use WHEN reviewing desktop, webview, worker, extension, RPC or process IPC boundaries for request/response schemas, error semantics, permission scope, versioning and caller coverage. Triggers: 'IPC contract', 'Tauri invoke', 'postMessage', 'bridge API', 'command schema', 'typed boundary'.
Use BEFORE introducing background jobs, queues, or scheduled tasks to choose delivery semantics, retry policy, and queue technology. Triggers: 'планировщик задач', 'cron', 'queue topology', 'фоновые задачи'.
Use WHEN building or reviewing LLM, prompt, agent, RAG, tool-use, routing, safety, or regression eval suites. Triggers: "LLM eval", "agent eval", "prompt regression", "golden corpus", "judge", "grader", "eval dataset".
Use WHEN designing, reviewing, or repairing LLM retrieval, RAG, embeddings, chunking, reranking, context packing, citation quality, memory retrieval, or graph-augmented context systems. Triggers: "RAG", "retrieval", "embeddings", "context quality", "rerank", "GraphRAG", "memory search".
Use WHEN a prototype, frontend handoff, or backend integration needs realistic mock data tied to API/schema contracts before backend is ready. Triggers: 'mock data contract', 'prepare mocks', 'frontend before backend', 'data-fed prototype', 'API fixtures'.
Use WHEN designing or reviewing model selection, local/hosted inference, latency and cost budgets, model rollout, fallback, prompt/model versioning, eval-to-release promotion, or AI production operations. Triggers: "model ops", "model selection", "inference", "latency", "cost", "fallback model", "LLM rollout".
Use WHEN diagnosing or planning work on routers, switches, Wi-Fi gateways, firewalls, NAT, VPN, DNS/DHCP, VLANs, routing stability, or ISP edge issues. Defaults to read-only diagnostics and requires explicit scoped approval before any network device mutation. Triggers: 'роутер', 'маршрутизатор', 'router config', 'network stabilization', 'VPN не работает', 'Wi-Fi падает'.
Use BEFORE shipping a service to production to design tracing, metrics, logs, SLOs, and on-call so incidents are detectable and debuggable. Triggers: 'логирование', 'метрики', 'трейсы', 'наблюдаемость'.
Use WHEN designing or reviewing payments, billing, subscriptions, entitlements, invoices, refunds, tax, reconciliation, webhooks, idempotency, and revenue-impacting release gates.
Use WHEN reviewing or improving performance to apply profile-first methodology with before/after benchmarks and root-cause bottleneck analysis. Triggers: 'оптимизация', 'медленно работает', 'profile', 'тормозит', 'ускорь'.
Use WHEN designing or reviewing privacy, PII handling, consent, retention, deletion, access logging, data minimization, cross-border data flow, and regulated-trust implementation contracts.
Use WHEN designing, reviewing, hardening, or debugging prompts, system instructions, agent prompts, tool-use policies, structured outputs, prompt evals, red-team suites, or user-intent interpretation. Triggers: 'prompt engineer', 'prompt architecture', 'system prompt', 'AI prompt', 'agent prompt', 'prompt injection', 'improve prompt', 'LLM instructions', 'промпт инженер', 'усиль промпт'.
Use WHEN preparing release, version bump, changelog, package surfaces, pre-push verification, rollback, release notes, or final 10/10 quality gates before merging or pushing to main.
Use WHEN auditing or planning security work to research CVE database, GitHub Security Advisories, and pattern-level vulnerabilities for project's stack. Triggers: 'security research', 'CVE на эту библиотеку', 'найди уязвимости', 'security audit'.
Use WHEN adding tracking/analytics to features to ensure event taxonomy, naming consistency, GDPR compliance, consent gating, and tracking plan documentation. Triggers: 'добавь аналитику', 'трекинг событий', 'GA4', 'настрой события', 'analytics'.
Use WHEN designing transactional or marketing email flows to ensure deliverability, accessibility, brand consistency, and lifecycle correctness. Triggers: 'email-флоу', 'lifecycle письма', 'transactional emails', 'настрой рассылку', 'welcome email'.
Use WHEN making product decisions (priority, scope, roadmap, OKR) at PM/CPO level for any user-facing feature or product area. Triggers: 'PRD', 'оформи продуктовое требование', 'product spec', 'приоритизируй', 'roadmap'.
Use WHEN designing test strategy or test suites to ensure coverage across test pyramid (unit, integration, e2e) with stack-appropriate patterns. Triggers: 'покрой тестами', 'напиши тесты', 'test plan', 'стратегия тестов', 'e2e сценарии'.
Use WHEN building or auditing public pages to ensure technical SEO (meta, schema.org, sitemaps, CWV) and content SEO (keyword targeting, structure, hreflang). Triggers: 'SEO аудит', 'оптимизируй SEO', 'мета-теги', 'schema.org', 'проверь SEO'.
Use WHEN converting vague requests into concrete contracts with acceptance criteria, edge cases, state machines, and system boundaries — READ-ONLY. Triggers: 'требования', 'requirements', 'формализуй', 'acceptance criteria', 'спецификация'.
Use WHEN implementing Android features in Jetpack Compose, Coroutines + Flow, Hilt DI, Room + WorkManager, Material 3 with Espresso + Compose UI tests. Triggers: 'android compose', 'jetpack', 'gradle config', 'kotlin coroutines'.
Use WHEN implementing ASP.NET Core features, controllers, minimal APIs, EF Core models, services, with xUnit tests and modern .NET patterns. Triggers: 'реализуй на ASP.NET', 'minimal API', 'EF Core модель', 'добавь controller .NET'.
Use WHEN designing Chrome MV3 extension architecture (manifest design, permissions strategy, service worker lifecycle, message-passing topology, content-script isolation, CSP, CWS publishing readiness) READ-ONLY. Triggers: 'chrome extension architecture', 'manifest v3 design', 'permission strategy', 'service worker design', 'mv2 to mv3 migration'. Triggers: 'спроектируй архитектуру расширения', 'manifest v3 архитектура', 'chrome extension архитектура', 'permission strategy', 'service worker дизайн'.
Use WHEN implementing Chrome MV3 extension features (popup, options page, side panel, content scripts, service worker, background events, message passing, storage) AFTER architecture is defined. Triggers: 'implement popup', 'write content script', 'service worker for extension', 'add side panel', 'message handler in extension', 'wire chrome.storage', 'inject content script'. Triggers: 'реализуй popup', 'напиши content script', 'service worker для расширения', 'добавь side panel', 'обработчик сообщений в расширении', 'подключи chrome.storage'.
Use WHEN designing Django application architecture (app boundaries, model graph, settings split, Celery + Channels topology, middleware ordering) READ-ONLY. Triggers: 'спроектируй Django архитектуру', 'границы приложений Django', 'topology для Celery', 'modular monolith на Django'.
Use WHEN implementing Django features (views/CBVs/DRF, ModelForms, signals, fixtures/factories, pytest-django) with disciplined ORM access. Triggers: 'реализуй фичу на Django', 'добавь view', 'напиши CBV', 'ModelForm для Django'.
Use WHEN designing or implementing DRF APIs (serializers, viewsets, permissions, pagination, filtering, throttling, simple-jwt) with N+1 discipline. Triggers: 'DRF serializer', 'viewset', 'permission для DRF', 'добавь endpoint на DRF'.
Use WHEN designing Elasticsearch/OpenSearch mappings, analyzers, sharding, ILM, search vs aggregation, fork-aware tradeoffs. Triggers: 'elasticsearch mapping', 'поисковый индекс', 'запросы elastic', 'opensearch'.
Use WHEN implementing Express.js APIs, middleware pipelines, route modules, validators, and error handlers with supertest coverage. Triggers: 'реализуй на Express', 'middleware на Express', 'добавь route', 'обработчик ошибок Express'.
Use WHEN implementing Fastify APIs, plugins, route schemas, hooks, decorators, error handlers, logging, graceful shutdown, and fastify.inject tests. Triggers: 'implement Fastify route', 'Fastify plugin', 'Fastify hook', 'Fastify schema', 'fastify.inject'.
Use WHEN designing FastAPI application architecture, dependency injection, async patterns, OpenAPI auto-gen, Alembic migrations READ-ONLY. Triggers: 'спроектируй FastAPI архитектуру', 'dependency injection FastAPI', 'topology для FastAPI', 'modular monolith на FastAPI'.
Use WHEN implementing FastAPI endpoints, models, services, async DB queries with pytest tests. Triggers: 'реализуй фичу на FastAPI', 'добавь endpoint FastAPI', 'async запрос FastAPI', 'Pydantic модель'.
Use WHEN implementing Flutter features, screens, BLoC/Riverpod state, platform channels, Dio API clients, with flutter_test + integration_test discipline. Triggers: 'flutter widget', 'state management', 'platform channel', 'riverpod'.
Use WHEN implementing Go HTTP services, handlers, repositories, workers with table-driven tests and idiomatic concurrency. Triggers: 'go микросервис', 'gin/echo handler', 'goroutine', 'go воркер'.
Use WHEN designing GraphQL schemas (schema-first vs code-first, federation v2, DataLoader, persisted queries, pagination, error handling, subscriptions, deprecation lifecycle) — cross-stack across Apollo, Hot Chocolate, Strawberry, gqlgen. Triggers: 'GraphQL схема', 'federation', 'resolver', 'dataloader'.
Use WHEN implementing iOS features in SwiftUI + Combine, async/await + actors, MVVM, App Intents, with XCTest + ViewInspector and accessibility discipline. Triggers: 'iOS экран', 'SwiftUI', 'Combine', 'swift async'.
Use WHEN designing or refining Eloquent models, relationships, scopes, casts to optimize queries and prevent N+1. Triggers: 'Eloquent модель', 'отношения Eloquent', 'migration', 'scope для модели'.
Use WHEN designing Laravel application architecture (modular monolith, DDD, Eloquent relationships, queue topology) READ-ONLY. Triggers: 'спроектируй Laravel архитектуру', 'modular monolith на Laravel', 'topology для Laravel', 'DDD в Laravel'.
Use WHEN implementing Laravel features, controllers, models, jobs, services with Pest tests and modern patterns. Triggers: 'реализуй фичу на Laravel', 'Eloquent', 'Artisan команда', 'middleware для Laravel'.
Use WHEN designing Laravel queue topology, jobs, retry strategies, Horizon configuration, idempotency, dead-letter handling, rate-limiting. Triggers: 'queue topology', 'Horizon', 'идемпотентность очередей', 'retry стратегия для job'.
Use WHEN designing MongoDB schema, indexes, sharding, aggregation pipelines, transactions, replica-set topology. Triggers: 'mongo schema', 'агрегации', 'sharding', 'mongodb индексы'.
Use WHEN designing MySQL/InnoDB schema, indexes, partitioning, replication, online DDL at scale. Triggers: 'mysql схема', 'innodb настройка', 'индексы mysql', 'репликация mysql'.
Use WHEN implementing NestJS modules, providers, controllers, guards, pipes, interceptors, repositories, and e2e tests with @nestjs/testing. Triggers: 'реализуй на NestJS', 'NestJS module', 'guard в NestJS', 'добавь controller NestJS'.
Use WHEN designing Next.js 14+ application architecture (server components default, streaming, ISR, edge runtime, route organization) READ-ONLY. Triggers: 'спроектируй Next.js архитектуру', 'server components топология', 'ISR стратегия', 'edge runtime дизайн'.
Use WHEN implementing Next.js 14+ pages, layouts, server actions, route handlers, mutations with TypeScript strict. Triggers: 'next.js страница', 'server action', 'API route', 'добавь layout Next.js'.
Use WHEN implementing Server Actions for mutations to enforce input validation, error handling, revalidation, and optimistic updates. Triggers: 'server action', 'Next.js server function', 'мутация через server action', 'revalidate в Next.js'.
Use WHEN designing Nuxt 3 application architecture (server vs client routing, Nitro engine choice, hybrid rendering SSR/SSG/ISR/CSR, Pinia stores, modules, layers, runtime config) READ-ONLY. Triggers: 'спроектируй Nuxt архитектуру', 'Nitro engine', 'topology для Nuxt', 'SSR vs SSG'.
Use WHEN implementing Nuxt 3 features (pages, layouts, middleware, server/api routes, useFetch + transform + key, useRuntimeConfig, useState SSR-aware, error.vue) with Vitest. Triggers: 'Nuxt страница', 'server route Nuxt', 'useFetch', 'middleware в Nuxt'.
Use WHEN designing Postgres schema, migrations, indexes, replication, partitioning at scale. Triggers: 'спроектируй postgres', 'индексы', 'миграция CONCURRENTLY', 'партицирование'.
Use WHEN deciding Rails architecture — Hotwire vs SPA, queue backend, ActionCable namespacing, engine boundaries, major architectural choices. Triggers: 'rails архитектура', 'engine', 'concern', 'hotwire vs spa'.
Use WHEN implementing Rails features — controllers, models, jobs, channels, Hotwire views with RSpec/Minitest and FormObject patterns. Triggers: 'rails контроллер', 'AR модель', 'миграция', 'sidekiq job'.
Use WHEN building standalone React (Vite/SWC) components requiring hooks-first patterns, state colocation, Suspense. Triggers: 'React компонент', 'добавь hook', 'Suspense', 'реализуй на React'.
Use WHEN designing Redis topology (single/Sentinel/Cluster), key schema, expiration policy, eviction, persistence, pub/sub vs streams, distributed locks. Triggers: 'redis topology', 'sentinel vs cluster', 'cache strategy', 'redis ключи'.
Use WHEN designing Spring Boot 3 / Spring 6 architecture, choosing WebFlux vs MVC, profiles, Actuator + observability, microservice boundaries READ-ONLY. Triggers: 'спроектируй Spring архитектуру', 'WebFlux vs MVC', 'topology для Spring', 'границы микросервисов'.
Use WHEN implementing Spring Boot features — REST controllers, services, JPA repositories, Bean Validation, Spring Security, Testcontainers integration tests. Triggers: 'реализуй на Spring', 'JPA репозиторий', 'добавь controller Spring', 'Spring Security настройка'.
Use WHEN implementing SvelteKit features, routes, hooks, form actions, load functions with Svelte 5 runes and Vitest/Playwright tests. Triggers: 'SvelteKit route', 'form action', 'load функция', 'добавь страницу SvelteKit'.
Use WHEN implementing or reviewing Tauri 2 Rust backend commands, plugins, window lifecycle, filesystem permissions, sidecars, updater flow, IPC boundaries, or desktop packaging. Triggers: 'Tauri Rust', 'src-tauri', 'tauri command', 'desktop IPC', 'sidecar', 'Rust backend'.
Use WHEN building Vue 3 components with Composition API, <script setup>, Pinia stores, typed props/emits, custom composables, Vitest + Vue Test Utils. Triggers: 'Vue компонент', 'composable', 'Pinia store', 'добавь component на Vue'.
Use WHEN stack changed (new modules, renamed files, removed files, new major dependencies) to sync agents/skills with new state. Triggers: 'sync проектные агенты', 'подтяни upstream', 'обнови агентов под новый стек', 'adapt'.
Use AFTER completing significant work (feature shipped, bug fixed, decision made, incident resolved) to add a memory entry capturing the learning for future agents. Triggers: 'добавь в память', 'save decision', 'сохрани решение', 'запиши learning'.
Use WHEN session starts in existing project OR WHEN agent reports stale-context to health-check artifacts: stale references, coverage gaps, weak artifacts, agent-freshness, override-rate, effectiveness signals. Triggers: 'health check', 'audit плагина', 'stale agents', 'аудит проекта'.
Use BEFORE implementing authentication TO choose the right flow (authorization-code+PKCE / client-credentials / device-code / resource-owner-password) for the use case. Triggers: 'auth flow', 'дизайн авторизации', 'OAuth выбор', 'как сделать логин'.
Use WHEN the user wants TO run a goal-until-complete autonomous multi-agent loop, epic, worktree run, or long-running session that turns a plan into tasks, dispatches specialists, supports status/resume/stop, and stops safely on policy, explicit budget, approval, or missing evidence. Triggers: 'autonomous loop', 'epic', 'worktree', 'goal-complete', 'эпик'.
Use BEFORE any creative work OR WHEN the user starts brainstorm, has finished brainstorm, or asks for the next planning step TO clarify intent, produce an approved spec, and hand off with Next: /supervibe-plan. Triggers: 'brainstorm', 'next step', 'брейншторм', 'брейншторм готов', 'я сделал брейншторм', 'план'.
Use WHEN starting new product OR major brand reset BEFORE production handoff to materialize an explicit design-system lifecycle at .supervibe/artifacts/prototypes/_design-system/. Candidate tokens guide review packets only; approved design-system state unlocks prototype work; final handoff metadata is stamped only after prototype approval. Triggers: 'нужен бренд', 'разработай бренд', 'фирстиль', 'брендбук', 'rebrand', 'design system', 'дизайн-система'.
Use WHEN browser-feedback system-reminder appears with click-region context AND active prototype, mockup, or presentation is open in preview server TO triage the comment, route to designer/deck/layout agent, and respond. Closes the user-to-browser-to-agent loop in real time. Triggers: 'browser-feedback received', 'кликнул на компонент', 'из браузера пришло', 'feedback overlay'.
Use WHEN setting up or modifying the build pipeline for a Chrome MV3 extension to choose bundler (Vite-CRXJS / WXT / Plasmo / vanilla), wire HMR for popup, generate icons, lint manifest, and produce a CWS-ready zip. Triggers: 'настрой сборку расширения', 'сделай build для chrome extension', 'добавь HMR в popup', 'упакуй для CWS', 'web-ext lint'.
Use BEFORE merging any change to systematically review code across 8 dimensions (correctness/security/readability/performance/coverage/error-handling/naming/docs) with severity ranking. Triggers: 'отревьюй код', 'code review', 'проверь PR', 'обзор кода'.
Use BEFORE making non-trivial changes to source code to find relevant existing code, similar patterns, and callers via hybrid keyword+semantic search. Triggers: 'найди код', 'кто вызывает', 'callers <symbol>', 'где используется'.
Use AFTER design-system-approved AND BEFORE prototype-handoff TO bridge brandbook tokens into a chosen component library (MUI, shadcn/ui, Radix UI, HeadlessUI, Mantine, or fully-custom). Decides which library fits, then generates the token bridge so the library renders with project palette/typography/motion.
Use BEFORE exiting any process skill that emits an artifact (requirements-spec, plan, agent-output, scaffold, prototype, research-output) to score it against its rubric and gate progression. Triggers: 'оцени по рубрике', 'confidence score', 'оцени уверенность', 'rubric check'.
Use WHEN design-facing agents need retrieval-backed style, UX, brand, deck, chart, collateral, or stack UI evidence TO ground decisions in project memory, code facts, approved design-system tokens, and the internal design intelligence data pack. Internal support only; no user-facing slash command.
Use WHEN facing 2+ independent tasks BEFORE starting them sequentially to determine if parallel subagent dispatch saves time without coordination cost. Triggers: 'parallel dispatch', 'fan-out', 'параллельно агентов', 'разнеси задачи'.
Use WHEN designing API error responses TO pick a consistent envelope (HTTP problem details problem+json / GraphQL union / gRPC status), define retry semantics, partial failure shapes. Triggers: 'error envelope', 'API ошибки формат', 'формат ошибок', 'problem+json'.
Use AFTER each agent task completion to track effectiveness (outcome, iterations, blockers, confidence-score, user-corrections) into effectiveness.jsonl. Triggers: 'оцени артефакт', 'log outcome', 'запиши результат', 'evaluate'.
Use WHEN an approved implementation-plan exists to execute it phase-by-phase with mandatory verification per task and confidence-gate per phase. Triggers: 'выполни план', 'execute plan', 'запусти план', 'погнали по плану'.
Use BEFORE running an A/B test or feature flag rollout to set up hypothesis, success metric, sample size calculation, randomization, and analysis plan. Triggers: 'a/b test', 'эксперимент', 'split-test', 'проверим гипотезу'.
Use BEFORE committing to any non-trivial decision (complexity ≥5) to enumerate ≥2 alternatives with tradeoffs and explicit chosen-rationale. Triggers: 'варианты', 'decision matrix', 'alternatives', 'сравни подходы'.
Use BEFORE shipping a risky feature TO design staged rollout (kill-switch / percentage / cohort), define rollback criteria, plan flag debt cleanup. Triggers: 'feature flag', 'staged rollout', 'постепенный релиз', 'фича-флаг'.
Use WHEN implementation complete and all verifications pass to decide how to integrate the work (merge, PR, archive, discard) with safety checks. Triggers: 'finish branch', 'merge готовое', 'закрой ветку', 'wrap up branch'.
Use WHEN bootstrapping host-aware Supervibe scaffold for a new or existing project to compose stack-pack into target with confidence-gate ≥9. Triggers: 'scaffolding', 'первичный setup', 'init проекта', 'genesis'.
Use WHEN production is broken (outage, data issue, security event) to triage, mitigate, root-cause, and write postmortem with timeline and action items. Triggers: 'incident', 'постмортем', 'упал прод', 'разбор инцидента'.
Use WHEN designing micro-interactions, animations, transitions, loading states to apply timing tiers, easing rules, and WOW-effect patterns from current 2026 design practice. Triggers: 'добавь анимацию', 'нужны переходы', 'микроинтеракция', 'оживи интерфейс'.
Use WHEN building a marketing or product landing page as a native HTML/CSS/JS prototype to scaffold with SEO meta, analytics hooks, copy review, accessibility, and explicit approval lifecycle for stack-agnostic handoff. Triggers: 'сделай лендинг', 'нужна landing страница', 'построй landing', 'дизайн посадочной'.
Use WHEN session starts OR WHEN user mentions visual/browser/desktop/data task to detect available MCP servers and proactively suggest agents that benefit from them. Triggers: 'найди MCP', 'какие tools доступны', 'есть MCP для', 'mcp discovery'.
Use WHEN a prototype, frontend handoff, or backend integration needs realistic mock data TO create schema-tied mock contracts and scenario fixtures before frontend implementation. Triggers: 'mock data contract', 'prepare mocks', 'data-fed prototype', 'frontend before backend'.
Use WHEN starting any new user-facing feature to orchestrate end-to-end flow from requirements to merge with all gates enforced. Triggers: 'новая фича', 'feature scaffold', 'давай новую фичу', 'начни фичу'.
Use BEFORE building any user-facing feature to write a Product Requirements Document framing problem, users, solution, success metrics, and out-of-scope. Triggers: 'PRD', 'продуктовая спецификация', 'product requirements', 'оформи PRD'.
Use BEFORE opening any PR or merging to main to run mandatory checks (typecheck, test, lint, dep audit, security scan) and capture evidence. Triggers: 'перед PR', 'pre-merge check', 'check before PR', 'пред-PR проверки'.
Use WHEN the user asks to create, revise, approve, or export a presentation deck to build a storyboard, HTML slide preview, feedback loop, approved PPTX, and Google Drive handoff. Trigger phrases: 'make a presentation', 'prepare pitch deck', 'create pptx', 'deck for investors', 'presentation mockup'.
Use AFTER generating HTML/CSS/JS mockup files TO spawn a local http://localhost preview server with hot-reload, share URL with user, optionally capture Playwright screenshot. Triggers: 'покажи в браузере', 'дай ссылку', 'открой превью', 'хочу потыкать'.
Use BEFORE starting any non-trivial task to search project memory for prior decisions, patterns, incidents, learnings, solutions matching current context. Triggers: 'поищи в памяти', 'past decisions', 'как мы делали раньше', 'project memory'.
Use AFTER a prototype is explicitly approved (.approval.json present with status=approved) to package it as a stack-agnostic handoff bundle at .supervibe/artifacts/prototypes/<slug>/handoff/ that any framework developer can pick up and promote to production. Triggers: 'утверди прототип', 'готов к разработке', 'передай разработчикам', 'handoff to dev', 'stack-agnostic export'.
Use WHEN user asks for design/mockup/UI exploration BEFORE implementing in framework to produce 1:1 capability-aware HTML/CSS/JS prototype in /prototypes for brand approval, feedback iteration, and framework-agnostic 1:1 transfer. Triggers: 'сделай мокап', 'покажи как будет выглядеть', 'нарисуй UI', 'нужен прототип', 'сделай макет'.
Use WHEN receiving code review feedback BEFORE implementing suggestions to evaluate each finding with technical rigor instead of performative agreement. Triggers: 'обработай review', 'feedback', 'комменты с PR', 'разбери ревью'.
Use BEFORE code-reviewer, PR, or AFTER a plan is written TO run a review loop with evidence, changed-file scope, plan risks, and next handoff. Triggers: 'pre-PR review', 'request review', 'готов к ревью', 'сделай ревью плана', 'review loop'.
Use BEFORE any new feature, bug fix, or refactor request to capture requirements with stack-aware questions and decide complexity routing (brainstorm vs plan vs exec). Triggers: 'requirements', 'intake', 'формализуй задачу', 'собери требования'.
Use WHEN validating, applying, or dry-running Supervibe project rules against a codebase or host artifact set TO prove the rules load, match intended files, produce actionable findings, and do not generate broad false positives.
Use WHEN rules-curator needs to detect contradictions/redundancy/gaps across host adapter rules OR after sync-rules. Triggers: 'аудит правил', 'rules check', 'проверь правила', 'rule audit'.
Use WHEN auditing or building public pages for SEO to verify technical SEO + content SEO using best-practices-researcher for current 2026 patterns. Triggers: 'SEO audit', 'мета-теги', 'проверь SEO', 'seo чеклист'.
Use WHEN session starts in unfamiliar project OR WHEN user mentions new stack to scan manifests, ask gap questions, build stack-fingerprint. Triggers: 'определи стек', 'stack discovery', 'какой тут стек', 'разведай проект'.
Use WHEN audit flagged weak/stale artifacts to deepen them from project context AND fresh research (consults research-agents for stale best-practices/dependency/security/infra/design). Triggers: 'усиль агента', 'strengthen <agent>', 'докрути артефакт', 'усилить'.
Use WHEN executing an implementation plan, epic, or atomic task wave with independent tasks AND subagents are available TO dispatch fresh subagent per task in isolated worktree/session with active session registry, heartbeat, stop/resume/status controls, and two-stage review. Trigger phrases: parallel subagents, fan-out, subagent-driven, atomic tasks, epic worktree, goal-until-complete autonomous, active session registry.
Use WHEN rules-curator updated rule in one project to propagate change to other projects of same stack (opt-in, with diff confirm). Triggers: 'синхронизируй правила', 'rules sync', 'разнеси правило', 'sync rules'.
Use WHEN encountering any bug, test failure, or unexpected behavior BEFORE proposing fixes to enforce hypothesis-evidence-isolation methodology. Triggers: 'дебаг', 'почему сломалось', 'найди корень', 'разбери баг'.
Use BEFORE writing any production code for a feature or bugfix to enforce red-green-refactor with integration-tests-first discipline. Triggers: 'TDD', 'red-green-refactor', 'failing test', 'тесты сначала'.
Use BEFORE designing tests for a new feature TO establish pyramid (unit/integration/e2e ratios), fixture isolation policy, flake budget, coverage triangulation strategy. Triggers: 'test strategy', 'покрытие тестами', 'тестовая пирамида', 'стратегия тестов'.
Use WHEN brandbook is approved AND frontend implementation needs theme to export brandbook tokens to framework-specific format (Tailwind/MUI/CSS vars/Style Dictionary). Triggers: 'выгрузи токены', 'export tokens', 'tailwind theme', 'нужна тема'.
Use WHEN a Supervibe trigger, command route, or skill handoff did not match, matched the wrong flow, or needs explanation TO diagnose intent, missing artifacts, confidence, blockers, and the next safe action.
Use AFTER a UI mockup, prototype, desktop/mobile shell, browser extension view, or implemented screen exists TO review layout, visual hierarchy, responsiveness, accessibility basics, interaction states, copy fit, and design-system adherence before handoff or implementation.
Use BEFORE feature work, autonomous session, or plan execution that needs isolation TO create or validate a git worktree, register active session ownership, keep the main workspace clean, heartbeat/status the session, and cleanup only after merge/PR. Trigger phrases: git worktree, isolated workspace, autonomous worktree, separate session, active session registry, heartbeat, cleanup.
Use BEFORE any claim of works/fixed/complete/passing/done to run a verification command and show its output as evidence — bans assertion without command output. Triggers: 'проверь', 'evidence', 'докажи что работает', 'верификация'.
Use AFTER an approved spec OR WHEN a plan is ready TO produce a phased implementation plan, require review loop, then split into atomic tasks and epic before execution. Triggers: 'составь план', 'сделал план', 'review plan', 'ревью луп', 'atomic', 'атомарные задачи', 'epic'.
Battle-tested Claude Code plugin for engineering teams — 50 agents, 188 skills, 68 legacy command shims, production-ready hooks, and selective install workflows evolved through continuous real-world use
Upstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.
Comprehensive startup business analysis with market sizing (TAM/SAM/SOM), financial modeling, team planning, and strategic research
Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Modifies files
Hook triggers on file write and edit operations
Uses power tools
Uses Bash, Write, or Edit tools
Share bugs, ideas, or general feedback.
Real-time statusline HUD for Claude Code - context health, tool activity, agent tracking, and todo progress
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claim