Skip to main content
Glama

Hyperstack Banner

A disciplined engineering harness that forces AI agents to use ground-truth docs, precise designs, and programmatic verification.


🚀 What is Hyperstack?

Hyperstack is a disciplined engineering harness for AI coding agents.

It provides the necessary Ground Truth (via 121 specialized MCP tools) and Adversarial Enforcement (via 29 rigorous skills) to transform a generic LLM into a high-precision Senior Engineer.

Unlike standard "polite" instructions, Hyperstack uses Iron Laws and a SessionStart hook to force agents to check real documentation, follow precise design specs, and provide binary verification before shipping.

Related MCP server: Unravel MCP

🚀 Installation

The fastest way to install Hyperstack is to let your AI agent do it for you. This works with Cursor, Windsurf, Roo Code, Claude Code, or Gemini. Simply paste this command:

Fetch and follow the instructions at https://raw.githubusercontent.com/orkait/hyperstack/main/install.md

The autopilot will autonomously detect your environment, install the MCP server (Docker or Local), and automatically link the Skills repository. This is the only step required.


2. Manual Configuration (Advanced)

If you aren't using the Autopilot, follow the Unified Bootstrap to set up both the MCP server and the Skills:

  1. Clone & Initialize:

    git clone https://github.com/orkait/hyperstack.git ~/.hyperstack
    cd ~/.hyperstack && bun install
  2. Run the Setup Script:

    bun scripts/setup.ts
  3. Follow the CLI: The script auto-detects your IDE, generates the JSON patch, and outputs the symlink command for your skill directory.

Supported platforms (verified April 2026, from official docs):

IDE / CLI

MCP Config Path

Skill Path

Schema

Claude Code

~/.claude.json

~/.claude/skills/hyperstack

JSON mcpServers

Gemini CLI

~/.gemini/settings.json

-

JSON mcpServers

Qwen Code

~/.qwen/settings.json

~/.qwen/skills/hyperstack

JSON mcpServers

Codex CLI

~/.codex/config.toml

-

TOML mcp_servers

Cursor

~/.cursor/mcp.json

.cursor/rules/ (project)

JSON mcpServers

superclaw

~/.config/superclaw/mcp.json

~/.config/superclaw/skills/hyperstack or superclaw plugin install --link

JSON mcpServers

Windsurf

~/.codeium/windsurf/mcp_config.json

-

JSON mcpServers

Kiro

~/.kiro/settings/mcp.json

-

JSON mcpServers

Zed

~/.config/zed/settings.json

-

JSON context_servers

VS Code + Copilot

~/.config/Code/User/mcp.json

.vscode/ (project)

JSON mcpServers

Roo Code

.roo/mcp.json (project)

.roo/rules/ (project)

JSON mcpServers

Cline

~/.config/Code/User/globalStorage/.../cline_mcp_settings.json

-

JSON mcpServers

Continue.dev

.continue/mcpServers/mcp.json (project)

-

JSON mcpServers

TIP

Runbun scripts/setup.ts and it will auto-detect your platform and output the exact patch - in the right format for your IDE. Codex users get TOML, everyone else gets JSON.

⚡ The Manifesto

It is not just a library; it is a disciplined harness made of three tightly-coupled layers:

  1. The Harness: Bootstraps the agent, routes internal specialist roles (hyper, website-builder), and enforces global development invariants.

  2. The MCP Ground Truth: 16 TypeScript plugins (121 tools) that provide deterministic data. No hallucinated imports or invented component specs.

  3. The Adversarial Gates: 29 skills with "Enforcement Teeth". These aren't suggestions; they are Iron Laws supported by rationalization tables that counter every excuse an agent uses to skip quality gates.

IMPORTANT

Use Hyperstack if you want to force your agent to check docs, write designs first, and prove its work. Skip Hyperstack if you want frictionless autocomplete. We prioritize intentional friction that catches bugs before they ship.

🏗️ Architecture

graph TD
    User([User Request]) --> Harness[Internal Harness]
    Harness --> Gates{Adversarial Gates}

    subgraph "The Enforcement Layer"
        Gates -->|Iron Laws| Skills[29 Skills with Enforcement Teeth]
        Skills -->|SessionStart| Injection[Auto-Context Injection]
    end

    subgraph "The Knowledge Layer"
        Skills -->|Ground Truth| MCP[16 MCP Plugins / 121 Tools]
        MCP -->|Designer| Design[DESIGN.md Pipeline]
        MCP -->|Verified| Code[Programmatic Compliance Check]
    end

    Code -->|Verification Evidence| User

    linkStyle 0 stroke:#9ca3af,stroke-width:2px
    linkStyle 1 stroke:#9ca3af,stroke-width:2px
    linkStyle 2 stroke:#9ca3af,stroke-width:2px
    linkStyle 3 stroke:#9ca3af,stroke-width:2px
    linkStyle 4 stroke:#9ca3af,stroke-width:2px
    linkStyle 5 stroke:#9ca3af,stroke-width:2px
    linkStyle 6 stroke:#9ca3af,stroke-width:2px
    linkStyle 7 stroke:#9ca3af,stroke-width:2px

🤖 Agent Model

If you are an agent reading this, your mental model must be: Process is Priority. You are operating within a system that requires deterministic grounding and formal design before implementation.

  • Conductance: hyper owners classification and gate verification.

  • Specialization: website-builder handles visual and frontend implementation.

  • Persistence: You run in a stable Docker container with a persistent lifecycle.


🚀 Quickstart

🐳 Docker (Default)

Hyperstack uses a persistent container plus docker exec. This keeps startup cheap across sessions and ensures 100% environment stability.

  1. Pull the image:

docker pull ghcr.io/orkait/hyperstack:main
  1. Start the persistent container:

docker rm -f hyperstack-mcp 2>/dev/null
docker run -d --name hyperstack-mcp --restart unless-stopped \
  --memory=512m --cpus=1 \
  --entrypoint sleep \
  ghcr.io/orkait/hyperstack:main infinity
  1. Add this to your MCP settings (~/.claude.json, Cursor, Windsurf, etc.):

{
  "mcpServers": {
    "hyperstack": {
      "command": "docker",
      "args": ["exec", "-i", "hyperstack-mcp", "bun", "/app/src/index.ts"]
    }
  }
}

🤖 Agent Autopilot

If you are using Claude Code, Cursor, Windsurf, Roo Code, or Gemini, you can use the autopilot to self-configure. Paste this at your agent:

Fetch and follow the instructions at https://raw.githubusercontent.com/orkait/hyperstack/main/install.md

The autopilot will detect your environment and propose the correct Docker-based configuration block.

🔧 Install the skills

The MCP server gives you tools. The skills give you discipline. Install both:

git clone https://github.com/orkait/hyperstack.git ~/.claude/skills/hyperstack

After installing, the SessionStart hook (at hooks/session-start.mjs) will auto-inject the hyperstack skill into every session. No manual activation needed.

💻 From source

git clone https://github.com/orkait/hyperstack.git
cd hyperstack
bun install
bun bin/hyperstack.mjs   # same entrypoint the published bin uses
bun start                # no build step
bun dev                  # watch mode
bun run build            # tsc --noEmit (type-check only, no dist output)

Node 18+ required.


🧠 The Four-Layer Framework

One framework, four layers: Ground Truth (MCP), Enforcement (Skills), Orchestration (Roles), and Personas (domain lenses, auto-engaged by trigger). Each persona binds its plugin, skills, and routing through one manifest.

Layer 1: MCP Plugins (Ground Truth)

Your AI calls these for deterministic data. Memory is not acceptable. Every plugin serves curated TypeScript data and architectural patterns.

Category

Plugins

Domain Coverage

🛠️ System

hyperstack

Autonomous Environment Detection, MCP Configuration Patching, Lifecycle

🎨 UI Engine

designer, design-tokens, ui-ux, shadcn

Design Systems, OKLCH, Typography, Accessibility, Component Specs

⚛️ Frontend

react, reactflow, motion, lenis

Next.js 15, RSC, Animation Curves, Smooth Scroll, DAG Layouts

🐹 Backend

echo, golang, rust

Professional Go Recipes, Rust Borrow Checker patterns, Clean Architecture

📈 Product & Growth

optimizer, product-manager, marketing, reflect

DSA technique matching, Product risk gates (JTBD, RICE), Positioning & GTM, Target-customer screen/feature review

TIP

121 tools across 16 plugins. Every tool is designed to provide the "Senior Engineer" answer, bypassing the "AI Slop" default.

Layer 2: Skills (Enforcement Teeth)

Markdown with adversarial enforcement. Each skill contains an Iron Law that the agent is bound to follow.

CAUTION

⚖️ The Iron Laws of Hyperstack

  • NO CODE without MCP grounding.

  • NO VISUAL CODE without an approved DESIGN.md.

  • NO COMPLETION CLAIMS without programmatic verification evidence.

  • NO REFACTOR without a failing test first.

  • NO PATTERN without a named Force.

These laws are backed by Rationalization Tables-pre-written counters to every excuse an AI agent uses to skip quality gates.

Layer 3: Agents (Orchestration & Routing)

The internal harness is what ties the public layers together by managing process and domains:

  • bootstrap is injected at session start from generated runtime context

  • hyper owns classification, routing, gates, and verification

  • website-builder specializes in website-facing design and implementation work

  • roles are internal and auto-called, not user-invoked commands

Layer 4: Personas (Domain Lenses)

Domain-expert lenses hyper auto-engages when their triggers match the request - the user never has to name them. Each binds an MCP plugin + skills + routing via personas/<id>/persona.json, and the trigger table is compiled into the session bootstrap from those manifests (single source of truth).

Persona

Mode

Lens

product-manager

gate

blocks net-new builds until value + viability risks carry evidence

bro

capability

generalist execution: decomposition, evidence discipline, code quality

marketing

capability

positioning, messaging, copy, brand voice, GTM

reflect

capability

reviews screens and features AS the real target customer (8 archetypes + panel mode)

Skill

Role

blueprint

Hard gate: no code without an approved design

forge-plan

MCP-verified task-by-task implementation plan

run-plan

Execute an existing plan

best-practices

8-step Senior SDE framework, 5 Iron Laws, patterns, coupling, review and ops rules

ship-gate

No completion claims without fresh verification evidence

deliver

Final verification and delivery

test-first

No production code without a failing test first

debug-discipline

Root cause first, 3-strike escalation

code-review

Dispatch reviewer subagent, handle feedback technically

autonomous-mode

Full end-to-end execution, only stops on failure

subagent-ops

Fresh agent per task, two-stage review

parallel-dispatch

Concurrent agent dispatch for independent tasks

worktree-isolation

Clean workspace isolation before feature work

Skill

Role

designer

Intention gate - produces DESIGN.md contract before any visual code

shadcn-expert

shadcn/ui Base UI architect - ONLY when user picks shadcn in designer Q11b

react-pro-coder

56 React/Next.js rules with stable IDs - rendering, state placement, audits

behaviour-analysis

UI/UX state audits, Nielsen heuristics, interaction matrices

security-review

OWASP audits, vulnerability checklists

readme-writer

Evidence-based README generation (this skill)

python-pro-coder

60 FastAPI/Pydantic v2/SQLAlchemy 2.0 rules - layering, schemas, async, dead deps

Skill

Role

hyperstack

Force-injected at session start via hook - the enforcement payload

testing-skills

RED-GREEN-REFACTOR pressure testing for skills using subagents

Full index at skills/INDEX.md. Regenerate with bun scripts/generate-skills-index.ts after adding or editing any skill.


🔒 Adversarial Enforcement

Ordinary skill markdown is a polite suggestion. Polite suggestion fails when an AI model is under pressure to "be helpful fast." Hyperstack skills are written adversarially:

  • 1% Rule: If there is even a 1% chance a skill applies, the agent must invoke it.

  • Rationalization Tables: We have already written down every excuse your AI will use to skip a gate, with a firm technical counter for each.

  • Loophole Closure: The "Spirit of the Law" is explicitly defined as the "Letter of the Law" to prevent shortcut-hunting.


🎨 The designer agent

When you say, “build me a SaaS dashboard”:

  1. SessionStart already puts in hyperstack, so AI know system is there.

  2. Blueprint skill sees visual job and sends it to hyperstack:designer.

  3. Designer skill runs designer_resolve_intent(product) to guess industry, personality, style, density, and mode.

  4. Designer asks 3 questions in base mode, or 12 questions in advanced mode.

  5. Like Q11b will ask what component library to use: shadcn, raw Tailwind, MUI, Mantine, Chakra, Ant Design, or custom.

  6. Designer makes a DESIGN.md contract with 10 parts: theme, colors, type, spacing, components, motion, elevation, do/don’ts, responsive rules, and anti-patterns.

  7. User approves the DESIGN.md.

  8. Forge-plan reads it and makes one task for each section. If user picked shadcn, it calls shadcn_get_component. If not, it builds from the DESIGN.md spec.

  9. Build tasks run with MCP tools as ground truth.

  10. designer_verify_implementation checks build against DESIGN.md.

  11. Ship-gate blocks final completion unless build passes the DESIGN.md rules.

AI cannot jump ahead. Every step has hard gate. Excuses already blocked by rationalization tables.


🛠️ Available Tools

📦 Supported Versions

Opinionated stack support for the latest stable stable releases. This prevents "AI Slop" by enforcing modern patterns (e.g., React 19 Actions, Tailwind v4 tokens).

Technology

Supported Version

Role

React

19.x

Core Library

Next.js

15.x

Application Framework

Tailwind CSS

v4.x

Design Tokens & Constraints

Motion

12.x (fka Framer Motion)

Orchestrated Animations

React Flow

12.x

Node-based Systems

Lenis

1.1.x+

Smooth Scroll Engine

Zustand

5.x

State Management

shadcn/ui

Base UI Edition

Component Primitive Patterns

Echo (Go)

v4.x

Backend Framework

Go

1.22+

Backend Language

Rust

1.77+

Systems Language

Bun

1.1.x+

Runtime Environment

  • designer_resolve_intent - Auto-detect industry, personality, style from product description

  • designer_list_personalities + designer_get_personality - 6 personality clusters from 58 real company design systems

  • designer_list_presets + designer_get_preset - 9 code-ready token presets (Linear, Stripe, Vercel, Apple, Carbon, shadcn, Notion, Supabase, Figma)

  • designer_get_industry_rules - 15 industry profiles with must-have/never-use constraints

  • designer_get_cognitive_law - 11 laws (Fitts, Hick, Miller, Gestalt, Von Restorff, Serial Position, F-Pattern, Z-Pattern, Jakob, Doherty, Peak-End)

  • designer_get_page_template - 13 page types with section anatomy

  • designer_get_composition_rules - Visual hierarchy, CRAP, whitespace, fold, reading patterns

  • designer_get_interaction_pattern - Form design, navigation, empty states, micro-interactions

  • designer_get_ux_writing - Button labels, error messages, confirmation dialogs

  • designer_get_landing_pattern - Hero, social proof, pricing, CTA optimization

  • designer_get_design_system - Specific values from 7 premium systems

  • designer_get_font_pairing - 21 curated pairings with Google Fonts imports

  • designer_get_anti_patterns - 50+ anti-patterns (the AI slop fingerprint) filterable by category/industry

  • designer_search - Cross-domain search

  • designer_generate_design_brief - Assemble structured brief

  • designer_generate_implementation_plan - Parse DESIGN.md into executable task list with MCP calls per section

  • designer_verify_implementation - Programmatic compliance check against DESIGN.md

Only invoked when the user explicitly chose shadcn in designer Q11b.

  • shadcn_get_rules - Architectural rules and mandatory checklist (call first)

  • shadcn_list_components - Curated component catalog

  • shadcn_get_component - Full spec: primitive, data-slots, variants, sizes

  • shadcn_get_snippet - Canonical usage example

  • shadcn_get_composition - Which components compose for a page type (bridge from designer page templates)

  • design_tokens_list_categories - 10 token categories (colors, spacing, grid, radius, shadows, motion, z-index, opacity, component sizing, typography)

  • design_tokens_get_category - CSS, rules, gotchas per category

  • design_tokens_get_color_ramp - OKLCH values + semantic roles

  • design_tokens_get_procedure - 8 step-by-step build procedures

  • design_tokens_get_gotchas - Aggregate implementation mistakes

  • design_tokens_generate - Complete Tailwind v4 token file from a palette

  • design_tokens_search - Cross-category search

  • ui_ux_list_principles - Browse by domain (typography, color, accessibility, responsive, motion)

  • ui_ux_get_principle - Rule, detail, CSS examples, anti-patterns

  • ui_ux_get_component_pattern - Button, card, badge, form specs

  • ui_ux_get_checklist - Pre-ship checklist per domain

  • ui_ux_get_gotchas - Common UI mistakes and fixes

  • ui_ux_search - Cross-domain search

  • reactflow_list_apis - Browse 56 APIs by kind

  • reactflow_get_api - Full reference: props, usage, tips

  • reactflow_search_docs - Full-text search

  • reactflow_get_examples - Curated code examples by category

  • reactflow_get_pattern - Enterprise patterns (zustand-store, drag-and-drop, SSR)

  • reactflow_get_template - Production-ready starters

  • reactflow_get_migration_guide - v11 to v12 breaking changes

  • reactflow_generate_flow - Generate a flow from prose

  • motion_list_apis - Browse 33 APIs

  • motion_get_api - Full reference with props and usage

  • motion_search_docs - Full-text search

  • motion_get_examples - Animation examples by category (gestures, scroll, layout)

  • motion_get_transitions - Transition reference for tween, spring, inertia

  • motion_generate_animation - Generate animation snippet from description

  • motion_cheatsheet - Quick reference

  • lenis_list_apis - Options, methods, events

  • lenis_get_api - Full reference with snippets

  • lenis_get_pattern - Next.js, GSAP, Framer Motion integrations

  • lenis_generate_setup - Complete Lenis setup

  • lenis_cheatsheet - Required CSS and pitfalls

  • lenis_search_docs - Full-text search

  • react_list_patterns - All React/Next.js patterns

  • react_get_pattern - Full implementation with anti-patterns

  • react_get_constraints - Hard rules (e.g., no useEffect for fetching)

  • react_search_docs - Search patterns and rules

  • echo_list_recipes - Browse 19 recipes

  • echo_get_recipe - Full recipe (jwt-auth, websocket, sse)

  • echo_list_middleware + echo_get_middleware - 13 middleware components

  • echo_decision_matrix - Echo vs standard library

  • echo_search_docs - Full-text search

  • golang_list_practices - Browse 18 best practices

  • golang_get_practice - Rule, reasoning, good/bad examples

  • golang_list_patterns + golang_get_pattern - 10 Go-idiomatic design patterns

  • golang_get_antipatterns - Common mistakes and fixes

  • golang_search_docs - Search practices and patterns

  • rust_list_practices + rust_get_practice - 18 best practices

  • rust_cheatsheet - Ownership rules, pointer types, performance

  • rust_search_docs - Search all practices


🤝 Contributing

We welcome contributions that follow the Disciplined Engineering standard.

  1. Plugins: Must follow the index.ts + data.ts + tools/ + snippets/ pattern.

  2. Skills: Must include category frontmatter and adhere to the Adversarial Enforcement style.

  3. Verification: All PRs must pass the full npm run build (Type-check) and CI suite.

# Regenerate the skills index after editing
bun scripts/generate-skills-index.ts

📄 License

MIT © Orkait | Adversarial philosophy inspired by Jesse Vincent's Superpowers.


🙏 Acknowledgements

The enforcement philosophy behind Hyperstack's gate skills - Iron Laws, 1% Rule, rationalization tables - was adopted from obra/superpowers (MIT © Jesse Vincent). We agreed with how it frames AI compliance: adversarially, not politely. See CREDITS.md.

Available Tools

118 tools
designer_generate_design_briefB

Generate a complete design brief for a product: resolves intent, then assembles visual theme, industry rules, style, cognitive laws, anti-patterns, design system inspiration, and composition rules

Inputs: product (required), mode, style, industry, +3 more.

designer_generate_implementation_planA

Parse a DESIGN.md into its 10 sections and generate a structured implementation plan. Each section becomes one or more tasks with exact MCP calls (shadcn_get_component, motion_generate_animation, design_tokens_generate, etc.) and self-review assertions. Caller reads DESIGN.md locally and passes the full text as design_md_content.

Inputs: design_md_content (required), framework.

designer_get_anti_patternsA

Get design anti-patterns filtered by category and/or industry

Inputs: category, industry.

designer_get_cognitive_lawA

Get a cognitive law: formula, key insight, UI applications, common violations, and academic source

Inputs: name (required).

designer_get_composition_rulesB

Get visual composition rules: key rule, detail, applications, and violations

Inputs: topic (required).

designer_get_design_systemB

Get design system reference: signature, key insights, typography, color, spacing, reference code, and cross-system convergences

Inputs: system (required).

designer_get_font_pairingA

Get curated font pairings filtered by mood (technical, elegant, friendly, editorial, bold, corporate, playful, luxury, startup, minimal) and/or industry. Returns heading + body + mono fonts with weights, tracking, line-height, Google Fonts import, and rationale. 21 pairings available.

Inputs: mood, industry.

designer_get_industry_rulesA

Get design rules for an industry: primary/secondary style, must-have features, never-use patterns, color mood, emotional target

Inputs: industry (required).

designer_get_interaction_patternC

Get interaction pattern: key rule, detail, best practices, and anti-patterns

Inputs: category (required).

designer_get_landing_patternB

Get landing page pattern: key stats, best practices, and anti-patterns

Inputs: topic (required).

designer_get_page_templateB

Get section anatomy, component inventory, and applicable cognitive laws for a page type. Covers landing, dashboard, auth, settings, checkout, blog, docs, admin, profile, error-page, ai-chat, pricing, and onboarding.

Inputs: type (required).

designer_get_personalityA

Get full personality profile: description, exemplars, visual vocabulary, mode, and CSS example

Inputs: cluster (required).

designer_get_presetA

Get a complete, code-ready design token preset based on a real premium design system. Returns colors (OKLCH), typography (font, scale, weights, tracking), spacing, radius, shadows, motion, and CSS example. Available presets: linear, stripe, vercel, apple, carbon, shadcn, notion, supabase, figma.

Inputs: name (required).

designer_get_ux_writingA

Get UX writing guideline: key rule, evidence, do/don't examples

Inputs: topic (required).

designer_list_personalitiesA

List all 6 designer personality clusters with descriptions and exemplar names

Inputs: no named parameters.

designer_list_presetsA

List all available design presets - complete, code-ready design token configurations based on real premium design systems (Linear, Stripe, Vercel, Apple, Carbon, shadcn, Notion, Supabase, Figma).

Inputs: no named parameters.

designer_resolve_intentB

Resolve a product description into a full design intent: industry, personality, style, mode, density, color mood, must-have/never-use lists

Inputs: product (required), userType, emotionalTarget.

designer_verify_implementationA

Verify implementation code against its DESIGN.md contract. Runs pattern checks for anti-patterns (AI purple, font-weight 500 everywhere, cold shadows, etc.), verifies OKLCH tokens present, checks for prefers-reduced-motion, and reports per-section compliance. Use this before ship-gate. Caller reads files locally and passes contents inline.

Inputs: code_files (required), design_md_content (required).

design_tokens_generateA

Generate CSS token scaffolding from a description. Returns ready-to-use CSS custom properties.

Inputs: description (required), framework.

design_tokens_get_categoryA

Get full details for a design token category including CSS examples, rules, and gotchas

Inputs: name (required).

design_tokens_get_color_rampA

Get a color ramp (brand/neutral/pop) with all 11 OKLCH stops, semantic roles, and light/dark mode usage

Inputs: name (required).

design_tokens_get_gotchasB

List all common design token mistakes and fixes across all categories

Inputs: no named parameters.

design_tokens_get_procedureA

Get the step-by-step token system build procedure. Steps 1-8 cover the full production workflow.

Inputs: step.

design_tokens_list_categoriesA

List all 10 design token categories with descriptions and architecture layer

Inputs: layer.

echo_decision_matrixB

Given a need or requirement, get the recommended Echo pattern and recipe to use.

Inputs: need.

echo_get_middlewareB

Get detailed config and usage for a specific Echo middleware.

Inputs: name (required).

echo_get_recipeA

Get a specific Echo framework recipe with full working Go code, gotchas, and related recipes.

Inputs: name (required).

echo_list_middlewareA

List all available Echo framework middleware with their purpose and recommended chain order.

Inputs: no named parameters.

echo_list_recipesA

List all Go Echo framework recipes. Optionally filter by category.

Inputs: category.

echo_search_docsB

Search Echo framework recipes and middleware by keyword.

Inputs: query (required).

golang_get_antipatternsB

List all Go anti-patterns to avoid

Inputs: no named parameters.

golang_get_patternB

Get a Go design pattern with idiomatic code

Inputs: name (required).

golang_get_practiceB

Get a Go best practice with good/bad code examples

Inputs: name (required).

golang_list_patternsA

List Go design patterns by category (creational, structural, behavioral, concurrency)

Inputs: category.

golang_list_practicesA

List Go best practices by topic and priority (P0=critical, P1=standard)

Inputs: topic, priority.

golang_search_docsA

Search Go best practices and design patterns by keyword

Inputs: query (required).

hyperstack_setupA

Identify current IDE/CLI environment and generate a tailored MCP configuration patch for Hyperstack.

Inputs: method, researchResult.

lenis_generate_setupB

Generate complete Lenis setup code from a natural-language description. Handles Next.js, GSAP, Framer Motion, basic React, and custom container scenarios.

Inputs: description (required).

lenis_get_apiA

Get detailed API reference for a specific Lenis API - props, options, usage examples, and tips.

Inputs: name (required).

lenis_get_patternA

Get a complete Lenis integration pattern with full production-ready code. Covers Next.js setup, GSAP integration, Framer Motion sync, custom containers, accessibility, and navigation.

Inputs: name (required).

lenis_list_apisA

List all Lenis smooth scroll APIs - ReactLenis component, useLenis hook, LenisRef and LenisOptions types.

Inputs: kind.

lenis_search_docsA

Search Lenis documentation by keyword. Searches API names, descriptions, code examples, and integration patterns.

Inputs: query (required).

marketing_briefA

Assemble a product-marketing brief for a specific brand: returns the ordered workflow (which frameworks/tools to apply, in sequence) for the deliverables you need. It routes the work - you produce the marketing.

Inputs: brand (required), deliverables.

marketing_get_anti_patternsA

Marketing red flags with fixes: feature-dumping, we-we copy, vague unproven claims, marketing-to-everyone, better-not-different, premature scaling, vanity metrics. Use to QA any marketing output before shipping.

Inputs: no named parameters.

marketing_get_awareness_stagesA

Eugene Schwartz's 5 stages of market awareness (unaware -> most-aware) + market sophistication, with how the copy must change per stage. Use to decide what to say before writing.

Inputs: no named parameters.

marketing_get_brand_strategyB

Evidence-based brand strategy: Byron Sharp's laws (penetration over loyalty, mental & physical availability, distinctive assets, double jeopardy), Binet & Field's 60/40 brand-vs-activation split, and category design (different-not-better). Use for the longer game.

Inputs: no named parameters.

marketing_get_channelsA

The 19 traction channels + the Bullseye framework (Traction, Weinberg & Mares): scan all channels, rank, test ~3 cheaply, focus on the one that works. Use to choose acquisition channels.

Inputs: no named parameters.

marketing_get_copywriting_formulasA

Copywriting formulas with fill-in templates: PAS, AIDA, BAB, FAB, PASTOR, 4 Ps, the 4 U's. Use to structure any headline, ad, landing page, or email. Apply the formula yourself - this supplies the templates.

Inputs: no named parameters.

marketing_get_growth_modelA

Growth model: AARRR pirate metrics (acquisition/activation/retention/referral/revenue) + growth loops vs funnels (Reforge) + the North Star metric. Use to design how growth compounds.

Inputs: no named parameters.

marketing_get_gtmA

Go-to-market motions (PLG vs sales-led vs marketing-led vs community-led, and when each fits) + the product-launch playbook (launch tiers and pre/launch/post phases). Use to decide how it reaches the market.

Inputs: no named parameters.

marketing_get_hooksA

Hook frameworks for the first 1-3 seconds (ads, short-form video, social, subject lines): curiosity/open-loop, pain-point, pattern-interrupt, contrarian, result, plus the 4-beat short-form structure and swipe openers. Use to grab attention before the scroll.

Inputs: no named parameters.

marketing_get_icpA

Define the Ideal Customer Profile: firmographics/psychographics, buying triggers, the 'who cares most' test, the early-vangelist profile, and the anti-ICP. Use during positioning to decide who EXACTLY it is for - not 'everyone'.

Inputs: no named parameters.

marketing_get_lifecycleA

Lifecycle / email marketing flows: welcome, onboarding, nurture, engagement/retention, win-back, and behavioral/transactional - triggered on behavior, not a blast calendar. Use to design the retention and monetization engine after acquisition.

Inputs: no named parameters.

marketing_get_messagingB

Messaging frameworks: value-proposition canvas (jobs/pains/gains), StoryBrand SB7 (customer is the hero), strategic narrative (Raskin - story = strategy), and the message hierarchy. Use after positioning is set.

Inputs: no named parameters.

marketing_get_persuasionA

Persuasion + headlines + landing-page anatomy: Ogilvy/Caples headline rules and swipe templates, Cialdini's 7 principles of influence, and the conversion-ordered landing page structure. Use while writing.

Inputs: no named parameters.

marketing_get_positioningA

Positioning framework (April Dunford): the 5 components in order (competitive alternatives -> unique attributes -> value -> target -> category) + the 3 positioning styles + a positioning-statement template. Use FIRST, before any messaging or copy.

Inputs: no named parameters.

marketing_get_pricingA

Pricing & packaging as a marketing lever: value-based pricing, good-better-best tiering, anchoring, fences between tiers, the value metric, and willingness-to-pay testing. Use to package and price the offer.

Inputs: no named parameters.

marketing_get_voiceA

Brand voice: the 12 archetypes (Mark & Pearson) with core desire/voice/example, plus the 4 tone dimensions (Nielsen Norman) for defining and applying a consistent voice. Use to set how the brand sounds.

Inputs: no named parameters.

marketing_get_voice_of_customerA

Voice-of-Customer / message mining (Copyhackers): how to harvest the customer's EXACT words from reviews, support, calls, and surveys, and turn them into copy. The method for getting marketing words straight from real customers. Use FIRST, before writing anything.

Inputs: no named parameters.

motion_generate_animationA

Generate a Motion for React animation snippet from a natural-language description. Returns ready-to-use JSX.

Inputs: description (required), elementType.

motion_get_apiA

Get detailed API reference for a specific Motion for React component, hook, or utility. Includes props, usage, examples, and tips.

Inputs: name (required).

motion_get_examplesC

Get code examples for a specific animation category

Inputs: category (required).

motion_get_transitionsA

Get the complete transition types reference (tween, spring, inertia, orchestration, per-value config)

Inputs: no named parameters.

motion_list_apisB

List all available Motion for React APIs (components, hooks, utilities)

Inputs: kind.

motion_search_docsA

Search Motion for React documentation by keyword. Searches API names, descriptions, and code examples.

Inputs: query (required).

optimizer_get_techniqueA

Get a technique's complexity, when it fits, the naive smell it replaces, and the web-search query to fetch its authoritative implementation. The catalog ships no code by design - algorithms are stable; verify the impl and exact complexity against a real source, not memory.

Inputs: name (required).

optimizer_list_classesA

List the algorithmic problem classes in the catalog with how many techniques each holds. The taxonomy to orient on before matching a problem.

Inputs: no named parameters.

optimizer_list_techniquesA

List the technique menu, optionally filtered by class. Each row is name + complexity + when to reach for it. No implementations - this is the recall aid.

Inputs: class.

optimizer_match_problemA

Match a problem description (or what the code does) to likely algorithmic classes and the candidate techniques to consider. Returns the menu to scan; the implementation is not shipped on purpose - web search the technique you pick.

Inputs: description (required).

product_manager_get_anti_patternsA

Codifiable PM red flags (Cagan, Doshi): feature-factory, reactivity, viability-avoidance, execution-misdiagnosis, opinion-requirement. Use to flag weak product reasoning.

Inputs: no named parameters.

product_manager_get_decision_toolsB

Decision-making toolkit: Type-1/Type-2 (reversible vs one-way-door) decisions, the 70% rule, pre-mortems, disagree-and-commit, decision records, and decision-vs-outcome quality. Use to make and document calls under uncertainty.

Inputs: no named parameters.

product_manager_get_discovery_rulesA

Continuous-discovery rules (Torres): weekly customer contact, never ask 'what do you want', elicit past-behaviour stories. Use before claiming a build is customer-grounded.

Inputs: no named parameters.

product_manager_get_four_risksA

The four product risks every build must address before shipping (Cagan/SVPG): value, usability, feasibility, viability - with which role owns each. The PM owns value and viability.

Inputs: no named parameters.

product_manager_get_jtbdA

Jobs-To-Be-Done framing (Christensen): a job is progress-in-context across functional/emotional/social dimensions, plus the Four Forces switching rule. Use to reframe a feature request as the underlying job.

Inputs: no named parameters.

product_manager_get_metricsC

Product metrics: North Star metric, One Metric That Matters, OKRs (outcomes not tasks), Google's HEART, AARRR, and input-vs-output / vanity-metric traps. Use to define what success is measured by.

Inputs: no named parameters.

product_manager_get_mvp_scopingA

MVP & scope-cutting: what an MVP really is, the Riskiest Assumption Test, vertical slicing / walking skeleton, story slicing, the cut test, and de-risk order (value->usability->feasibility). Use to cut a build to its thinnest valuable slice.

Inputs: no named parameters.

product_manager_get_prioritizationB

Prioritization frameworks beyond RICE: MoSCoW (release scoping), Kano (basic/performance/excitement), ICE, WSJF / cost-of-delay, and opportunity scoring (importance vs satisfaction gap). Use to decide what to build and in what order.

Inputs: no named parameters.

product_manager_get_product_senseB

How to develop product sense (Jules Walter, Julie Zhuo): customer immersion, product teardowns/critiques, principles-first learning, and closing the predict/outcome loop. The deliberate practice for building product judgment.

Inputs: no named parameters.

product_manager_get_roadmappingA

Roadmapping: Now/Next/Later horizons, outcome-based (not feature-based) roadmaps, theme-based organization, and GIST. Use to communicate direction without faking date precision or sliding into a feature factory.

Inputs: no named parameters.

product_manager_get_strategy_rulesB

Product-strategy rules (Cagan): focus on 2-3 levers, saying no is the act of prioritization, a long list is a non-strategy. Use to cut scope.

Inputs: no named parameters.

product_manager_opportunity_vs_solutionA

Returns the Torres opportunity-vs-solution rubric (the test + reframe examples) for YOU to apply - it does not classify for you. Use to reframe a solution-shaped request into the underlying need.

Inputs: statement.

product_manager_resolve_product_decisionA

The PM gate. Supply YOUR value and viability assessments for the proposal; returns a real verdict - PASS (both addressed), BLOCK (net-new build with a PM-owned risk unaddressed), or ADVISORY (tweak). PASS is earned by actually assessing value AND viability, not automatic. The judgment is yours; this enforces that you made it.

Inputs: description (required), isNetNew, valueAssessment, viabilityAssessment.

product_manager_score_riceA

Compute a RICE prioritization score = (Reach x Impact x Confidence) / Effort (Intercom). Confidence is 0-1. Use to rank competing initiatives.

Inputs: reach (required), effort (required), impact (required), confidence (required).

product_manager_validate_job_statementA

Returns the 7-point JTBD job-statement criteria (Christensen/Ulwick) for YOU to judge a statement against - it does not pass/fail for you. Use to sharpen a vague job into one with context + progress.

Inputs: statement.

reactflow_generate_flowB

Generate a React Flow component from a natural-language description. Returns ready-to-use TSX with proper imports.

Inputs: description (required), controlled.

reactflow_get_apiA

Get detailed API reference for a specific React Flow component, hook, utility, or type. Includes props, usage, examples, and tips.

Inputs: name (required).

reactflow_get_examplesC

Get code examples for a specific React Flow category

Inputs: category (required).

reactflow_get_migration_guideA

Get the React Flow v11 to v12 migration guide with all breaking changes, import changes, and type changes

Inputs: no named parameters.

reactflow_get_patternB

Get an enterprise React Flow pattern with full implementation code. Patterns include store architecture, undo/redo, drag-and-drop, auto-layout, context menus, copy/paste, save/restore, DAG validation, keyboard shortcuts, performance, dark mode, SSR, subflows, edge reconnection, and more.

Inputs: pattern (required).

reactflow_get_templateA

Get a production-ready code template: custom-node (Tailwind + toolbar + handles + status), custom-edge (delete button + BaseEdge), or zustand-store (full store with selectors)

Inputs: template (required).

reactflow_list_apisA

List all React Flow v12 APIs - components, hooks, utilities, and types

Inputs: kind.

reactflow_search_docsA

Search React Flow documentation by keyword. Searches API names, descriptions, code examples, and tips.

Inputs: query (required).

react_get_constraintsA

List all forbidden React/Next.js patterns and their reasons

Inputs: no named parameters.

react_get_patternA

Get a React/Next.js pattern with full code example and anti-pattern

Inputs: name (required).

react_list_patternsA

List all React/Next.js patterns by category

Inputs: category.

react_search_docsB

Search React/Next.js patterns and constraints by keyword

Inputs: query (required).

reflect_get_panelA

Get the full multi-persona panel for reviewing a screen from every relevant lens at once: the panel protocol (relevance-based assembly, own-voice rules, collision naming, blocker vetoes), the shared voice rules, and the full lens docs. Pass ids to scope the panel; omit for all archetypes.

Inputs: ids.

reflect_get_personaA

Get one Reflect reviewer archetype's full lens, what they catch/undervalue, and voice - so you can review a screen AS them. Then apply the shared voice rules (reflect_get_voice_rules).

Inputs: id (required).

reflect_get_voice_rulesA

The shared Reflect voice contract: sounds-human vs sounds-like-AI, the 5 moods, the 6 in-head checks, how to reply, plus sample replies. Apply to every archetype so the review sounds like a real person, not a UX bot.

Inputs: no named parameters.

reflect_list_personasA

List the Reflect reviewer archetypes (the target-customer lenses you can review a screen as). Default is Morgan. Switch with reflect_get_persona(id).

Inputs: no named parameters.

rust_cheatsheetA

Quick Rust reference: ownership rules, error handling, clippy commands, and key patterns

Inputs: no named parameters.

rust_get_practiceB

Get a Rust best practice with code examples

Inputs: name (required).

rust_list_practicesC

List Rust best practices by chapter

Inputs: chapter.

rust_search_docsB

Search Rust best practices by keyword

Inputs: query (required).

shadcn_get_componentA

Get full details for a shadcn/ui component: base primitive, data-slots, variants, sizes, usage example, and which other components it pairs with. Uses curated reference data.

Inputs: name (required).

shadcn_get_compositionA

Get shadcn/ui component composition for a specific page type. Returns which components to combine per section with rationale. This is the bridge from designer's page templates to concrete implementation. Use after designer_get_page_template.

Inputs: page_type (required).

shadcn_get_rulesA

Get the architectural rules and mandatory checklist for shadcn/ui (Base UI edition). Call this before proposing any new component or modification.

Inputs: no named parameters.

shadcn_get_snippetA

Get a usage code snippet for a shadcn/ui component. Returns the canonical example showing variants, sizes, and composition patterns.

Inputs: name (required).

shadcn_list_componentsA

List all curated shadcn/ui components (Base UI edition). Optionally filter by category: button, input, card, dialog, dropdown, tabs, table, form, navigation, feedback, overlay, data-display.

Inputs: category.

ui_ux_get_checklistA

Get quality checklist for a UI/UX domain before shipping

Inputs: domain (required).

ui_ux_get_component_patternB

Get component pattern spec including variants, states, and sizing rules

Inputs: name (required).

ui_ux_get_gotchasA

List all common UI/UX mistakes and fixes, optionally filtered by domain

Inputs: domain.

ui_ux_get_principleA

Get full details for a UI/UX principle including examples, anti-patterns, and CSS examples

Inputs: name (required).

ui_ux_list_principlesA

List all UI/UX principles by domain (typography, color, spacing, elevation, motion, accessibility, responsive, components)

Inputs: domain.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    B3.4/5.0

    Scored across 118 tools

    Disambiguation4/5

    Domain prefixes (reactflow_, marketing_, etc.) clearly separate tool groups, but the high number (118) across many domains may still cause confusion, especially with similar verb patterns like 'get_*' across different domains. Within a domain, tools are mostly distinct.

    Naming Consistency5/5

    All tools follow a consistent <domain>_<verb>_<noun> pattern (e.g., reactflow_list_apis, marketing_get_positioning). Minor exceptions like hyperstack_setup don't break the overall consistency.

    Tool Count2/5

    118 tools is excessive for a single MCP server. While each sub-domain individually might warrant its count, the aggregate scope is too broad and will overwhelm agent prompts. The server tries to cover too many unrelated areas (e.g., React Flow, marketing, product management, etc.).

    Completeness3/5

    Within each sub-domain, the tool surface is fairly complete for knowledge retrieval (list, get, search, generate). However, the server is a collection of static knowledge bases with little cross-domain integration or action-oriented tools, leaving gaps in workflows that require combining domains.

    Maintenance

    ActivityStale
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      D
      maintenance
      A methodology and MCP server for agent-driven software development where humans write specs and agents implement code, enforced by six mechanical gates to ensure spec validity, contracts, tests, and review.
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      A deterministic AST evidence engine that forces AI agents to debug using verified execution facts instead of pattern-matching symptoms, enabling hallucination-free debugging for MCP-compatible agents.
      27 npm
      Business Source 1.1
    • A
      license
      A
      quality
      D
      maintenance
      An MCP server that enforces hard quality gates on AI coding agents, including forced web search, planning, and empirical test execution, to prevent common failure modes and ensure production-grade code.
      13
      30
      MIT