anchi-kit 1.2.2 → 1.2.3

package/.antigravity/skills/_packs/common/pack-architecture.md

@@ -0,0 +1,40 @@
+# Pack: Architecture
+
+Tier: Core
+Load Cost: High
+Keywords: ddd, solid, patterns, planning, system-design
+
+## Scope
+
+This pack covers **Software Architecture and High-Level Planning**.
+
+- **Includes**: Domain-Driven Design (DDD), Modular Monolith, SOLID principles, Project Structure, Implementation Planning.
+- **Excludes**: Specific coding syntax.
+
+## Canonical Patterns (Best Practices)
+
+### 1. Structure (Modular Monolith)
+
+- **Feature Modules**: Group by feature (`features/billing/`), not type (`controllers/`).
+- **Encapsulation**: Modules expose a Public API (`index.ts`); internals are private.
+- **Dependency Rule**: Outer layers depend on inner layers (Core Domain).
+
+### 2. SOLID Principles
+
+- **SRP**: Single Responsibility per class/module.
+- **DIP**: Depend on abstractions (interfaces), not concretions.
+
+### 3. Planning
+
+- **Implementation Plan**: Always create `implementation_plan.md` before coding complex features.
+- **Task Breakdown**: Break tasks into < 1 hour chunks.
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Big Ball of Mud**: Circular dependencies between all modules.
+- ❌ **God Objects**: Classes that do everything.
+- ❌ **Coding without Plan**: Jumping into code for complex tasks.
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded Architecture Pack."

package/.antigravity/skills/_packs/common/pack-devops.md

@@ -0,0 +1,43 @@
+# Pack: DevOps
+
+Tier: Domain
+Load Cost: Medium
+Keywords: docker, ci-cd, git, deployment, infrastructure, automation
+
+## Scope
+
+This pack covers **Development Operations, Infrastructure, and Automation**.
+
+- **Includes**: Git workflow, CI/CD pipelines (GitHub Actions), Docker containers, Deployment scripts, MCP Server management.
+- **Excludes**: Application code logic (Frontend/Backend).
+
+## Canonical Patterns (Best Practices)
+
+### 1. Git Workflow
+
+- **Conventional Commits**: `feat:`, `fix:`, `chore:`, `docs:`.
+- **Branching**: `main` (stable), `develop` (integration), `feature/xyz`.
+- **Clean History**: Rebase before merge preferred over merge commits for cleaner history.
+
+### 2. CI/CD (GitHub Actions)
+
+- **Fast Feedback**: Run lint and tests on every PR.
+- **Cache Dependencies**: Use `actions/cache` for npm/pnpm.
+- **Secret Management**: Use GitHub Secrets, never hardcode tokens in YAML.
+
+### 3. MCP (Model Context Protocol)
+
+- **Builder**: Use standard MCP SDKs.
+- **Management**: Register servers via config files, not ad-hoc commands.
+- **Security**: MCP servers run locally; trust bounds apply.
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Commit "WIP"**: Avoid vague messages. Squash WIP commits before merge.
+- ❌ **Broken Build**: Never merge code that fails CI.
+- ❌ **Manual Deploys**: Automate via pipeline whenever possible.
+- ❌ **Root Keys**: Never use root access keys in CI; use scoped service accounts.
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded DevOps Pack."

package/.antigravity/skills/_packs/common/pack-productivity.md

@@ -0,0 +1,37 @@
+# Pack: Productivity
+
+Tier: Core
+Load Cost: Low
+Keywords: files, research, data, docs, organization
+
+## Scope
+
+This pack covers **General Productivity, File Management, and Data Handling**.
+
+- **Includes**: File searching, filtering, documentation organization, data analysis (CSV/JSON), learning patterns.
+
+## Canonical Patterns (Best Practices)
+
+### 1. File Management
+
+- **Search**: Use `scout` or `fd` to find files efficiently.
+- **Organization**: Keep directories flat where possible; group by domain.
+
+### 2. Documentation
+
+- **Proximity**: Keep docs close to code (`README.md` in folders).
+- **Updates**: Update docs _with_ code changes in the same PR.
+
+### 3. Data Analysis
+
+- **Validation**: Validate input data schema.
+- **Visualization**: Summarize data trends clearly.
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Stale Docs**: Documentation that contradicts code.
+- ❌ **Messy Workspace**: files scattered in root.
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded Productivity Pack."

package/.antigravity/skills/_packs/common/pack-quality.md

@@ -0,0 +1,41 @@
+# Pack: Quality
+
+Tier: Domain
+Load Cost: Medium
+Keywords: testing, debug, review, security, performance, kaizen
+
+## Scope
+
+This pack covers **Quality Assurance, Testing, and Code Review**.
+
+- **Includes**: Unit Testing (Vitest/Jest), E2E (Playwright), Debugging strategies, Security audits, Performance profiling.
+- **Excludes**: Implementation details.
+
+## Canonical Patterns (Best Practices)
+
+### 1. Testing (TDD)
+
+- **Red-Green-Refactor**: Write test -> Fail -> Pass -> Clean.
+- **AAA Pattern**: Arrange, Act, Assert.
+- **Mocking**: Mock external boundaries (API, DB), test logic in isolation.
+
+### 2. Review
+
+- **Readability**: Code should be self-documenting.
+- **Maintainability**: Avoid complex abstractions early.
+- **Security**: Sanitize inputs, check authorization.
+
+### 3. Debugging
+
+- **Systematic Approach**: Reproduce -> Isolate -> Fix -> Verify.
+- **Logs**: Use structured logging.
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Testing Implementation**: Test behavior, not internal state.
+- ❌ **Flaky Tests**: Tests relying on timing/network without mocks.
+- ❌ **Premature Optimization**: Optimize only after profiling.
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded Quality Pack."

package/.antigravity/skills/_packs/data/pack-ai.md

@@ -0,0 +1,41 @@
+# Pack: AI (Cognitive)
+
+Tier: Core
+Load Cost: Medium
+Keywords: llm, prompt, mcp, agents, reasoning, research
+
+## Scope
+
+This pack covers **AI Capabilities, Reasoning Strategies, and Research**.
+
+- **Includes**: Prompt Engineering, Agent Personas, Sequential Thinking, Research methodologies, Context Management.
+- **Excludes**: Specific implementation details (coding).
+
+## Canonical Patterns (Best Practices)
+
+### 1. Sequential Thinking
+
+- **Step-by-Step**: Break complex problems into atomic steps.
+- **Self-Correction**: Review previous steps before proceeding.
+- **Hypothesis Testing**: Formulate hypotheses and verifying them effectively.
+
+### 2. Research (Scout)
+
+- **Breadth-First**: Scan wide before diving deep.
+- **Source Verification**: Prefer official docs and code over generic articles.
+- **Synthesis**: Combine multiple sources into a coherent answer.
+
+### 3. Multi-Modal
+
+- **Images**: Analyze UI screenshots for design fidelity.
+- **Context**: Provide relevant context files, not whole folders.
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Hallucination**: Inventing APIs or libraries. Always verify.
+- ❌ **Looping**: Getting stuck in "I will fix" loops without changing strategy.
+- ❌ **Ignoring Constraints**: Forgetting project-specific rules (e.g., "No Tailwind").
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded AI Pack."

package/.antigravity/skills/_packs/data/pack-data-science.md

@@ -0,0 +1,36 @@
+# Pack: Data Science
+
+Tier: Domain
+Load Cost: High
+Keywords: python, pandas, numpy, visualization, analysis, machine-learning, plot
+
+## Scope
+
+This pack covers **Deep Data Science and Analysis**.
+
+- **Includes**: Data manipulation (Pandas/Polars), Visualization (Matplotlib/Seaborn), Statistical Analysis, Basic ML (Scikit-learn).
+- **Primary Lang**: Python (assumed execution environment or script generation).
+
+## Canonical Patterns (Best Practices)
+
+### 1. Data Pipeline
+
+- **ETL**: Extract -> Transform -> Load. Clean data before analysis.
+- **Vectorization**: Use Pandas/NumPy vector operations, avoid Python loops.
+- **Reproducibility**: Use Jupyter Notebooks or scripts with seed fixing.
+
+### 2. Visualization
+
+- **Clarity**: Label axes, titles, and legends.
+- **Color**: Use accessible color palettes.
+- **Context**: Annotate outliers or key events.
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Mutating Data**: Modifying DataFrames in-place implicitly.
+- ❌ **Magic Numbers**: Hardcoding boundaries without explanation.
+- ❌ **Overfitting**: training on test data.
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded Data Science Pack."

package/.antigravity/skills/_packs/mobile/pack-mobile.md

@@ -0,0 +1,40 @@
+# Pack: Mobile
+
+Tier: Domain
+Load Cost: Medium
+Keywords: react-native, expo, ios, android, mobile
+
+## Scope
+
+This pack covers **Mobile App Development** using React Native and Expo.
+
+- **Includes**: Native Components, Navigation (React Navigation/Expo Router), native modules, platform-specific code.
+- **Excludes**: Pure web components (DOM-based).
+
+## Canonical Patterns (Best Practices)
+
+### 1. Framework
+
+- **Expo Framework**: Preferred for new projects (File-based routing, Prebuild).
+- **NativeWind**: Tailwind CSS for React Native.
+
+### 2. Components
+
+- **Core Components**: Use `<View>`, `<Text>`, `<Pressable>` (not `<div>`, `<span>`).
+- **Lists**: Use `<FlatList>` or `<FlashList>` for long lists (virtualization).
+- **Images**: Use `expo-image` for caching and performance.
+
+### 3. Navigation
+
+- **Architecture**: Stack Navigator for flows, Tab Navigator for main shell.
+- **Deep Linking**: Configure schemes for external access.
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Web-only APIs**: Using `window`, `document`, or `localStorage` (Use `AsyncStorage`).
+- ❌ **Blocking JS Thread**: Heavy computations on UI thread. Use Worklets (Reanimated).
+- ❌ **Unoptimized Images**: Loading huge raw images.
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded Mobile Pack."

package/.antigravity/skills/_packs/web/pack-backend.md

@@ -0,0 +1,61 @@
+# Pack: Backend
+
+Tier: Domain
+Load Cost: High
+Keywords: node, api, database, auth, payment, backend, server
+
+## Scope
+
+This pack covers **Backend Development** logic, API design, and Data Management.
+
+- **Includes**: API endpoints (Next.js App Router), Database interactions (Prisma/Drizzle), Authentication (Auth.js), Payments (Stripe), Business Logic.
+- **Excludes**: UI components, detailed DevOps/Infrastructure (see `pack-devops`).
+
+## Canonical Patterns (Best Practices)
+
+### 1. API Design
+
+- **Resource Oriented**: `GET /api/users`, `POST /api/users`.
+- **Validation**: Use Zod for all request body/param validation.
+- **Error Handling**: Return structured errors `{ code, message, details }`.
+- **Type Safety**: Share types between Backend and Frontend (Monorepo/RPC).
+
+### 2. Database Interactions
+
+- **ORM**: Use Prisma or Drizzle.
+- **Migrations**: Always verify migrations locally before commit.
+- **Indexing**: Add indexes for foreign keys and frequently searched columns.
+- **Connection Pooling**: Use connection pooling for Serverless environments.
+
+### 3. Authentication & Security
+
+- **Auth.js / Better-Auth**: Standard implementations preferred.
+- **Middleware**: Protect routes via middleware, not just client-side checks.
+- **RBAC**: Implement Role-Based Access Control logic in services.
+- **Secrets**: Never commit `.env`. Access via `process.env`.
+
+### 4. Payments (Stripe)
+
+- **Webhooks**: Always verify webhook signatures securely.
+- **Idempotency**: Handle duplicate webhook events gracefully.
+- **Client/Server Split**: Create PaymentIntent on server, confirm on client.
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Logic in Controllers**: Keep Route Handlers thin; move logic to Services/UseCases.
+- ❌ **N+1 Queries**: Loop queries inside loops. Use `include` or `Promise.all`.
+- ❌ **Trusting Client**: Never trust `req.body` without Zod validation.
+- ❌ **SQL Injection**: Never concat strings for queries. Use ORM or parametrized queries.
+- ❌ **Hardcoded IDs**: Do not rely on specific DB IDs in logic.
+
+## Specialized Modules
+
+### Shopify Integration
+
+- Use Shopify Admin API / Storefront API.
+- Respect rate limits.
+- Handle webhook HMAC verification strict.
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded Backend Pack."

package/.antigravity/skills/_packs/web/pack-frontend.md

@@ -0,0 +1,66 @@
+# Pack: Frontend
+
+Tier: Domain
+Load Cost: High
+Keywords: react, nextjs, tailwind, ui, ux, components, frontend
+
+## Scope
+
+This pack covers all aspects of **Frontend Development** within the Anchi-Kit ecosystem.
+
+- **Includes**: Component architecture, UI styling (Tailwind), State Management, React Hooks, Next.js Routing, Client-Side logic, Accessibility (a11y), Animations.
+- **Excludes**: Backend API logic (see `pack-backend`), Database schemas, DevOps pipelines.
+
+## Canonical Patterns (Best Practices)
+
+### 1. Component Design
+
+- **Single Responsibility**: Each component should do one thing well.
+- **Composition over Inheritance**: Use children props and slots.
+- **Props Interface**: Always define explicit interfaces for props (TypeScript).
+- **Atomic Design**: Structure components as Atoms, Molecules, Organisms, Templates.
+
+### 2. Styling (Tailwind CSS)
+
+- **Utility-First**: Use Tailwind classes directly in markup.
+- **Tokens**: Use `theme('colors...')` for consistency.
+- **Responsive**: Mobile-first design (`min-width` default, `md:`, `lg:` overrides).
+- **ClassName Merging**: Use `cn()` utility (clsx + tailwind-merge) for conditional classes.
+
+### 3. State Management
+
+- **Local State**: `useState` for simple UI toggles.
+- **Server State**: Use `useQuery` or Next.js Server Components for data fetching.
+- **Global State**: Minimal use of Context/Zustand only when absolutely necessary (e.g., Theme, Auth User).
+
+### 4. Performance
+
+- **Image Optimization**: Use `next/image`.
+- **Lazy Loading**: Use `dynamic()` for heavy components below the fold.
+- **Memoization**: `useMemo`/`useCallback` only when profiling shows re-render issues.
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Hardcoded Styles**: Avoid `style={{ margin: 10 }}`. Use Tailwind classes.
+- ❌ **Prop Drilling**: Passing data through >3 layers. Use Composition or Context.
+- ❌ **Client-Side Heavy**: Making everything `"use client"`. Default to Server Components.
+- ❌ **Magic Numbers**: Using arbitrary pixel values (e.g., `width: 317px`). Use design tokens.
+- ❌ **Huge Components**: Files > 300 lines should likely be split.
+
+## Specialized Modules
+
+### 3D / WebGL (Three.js)
+
+- Use `@react-three/fiber`.
+- Keep canvas logic separate from DOM UI.
+- optimize models (GLB/GLTF) via `gltfjsx`.
+
+### Mobile (React Native / Expo)
+
+- Use `StyleSheet.create` or strict Tailwind integration (NativeWind).
+- Handle platform differences (`Platform.OS`).
+- Avoid web-only APIs is critical paths.
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded Frontend Pack."

package/.antigravity/skills/_packs/web3/pack-blockchain.md

@@ -0,0 +1,37 @@
+# Pack: Blockchain
+
+Tier: Domain
+Load Cost: High
+Keywords: blockchain, web3, solidity, smart-contract, ethereum, crypto, wallet
+
+## Scope
+
+This pack covers **Web3 and Blockchain Development**.
+
+- **Includes**: Smart Contract development (Solidity), Interaction (Ethers.js/Viem), Testing (Hardhat/Foundry), Security (Reentrancy prevention).
+- **Excludes**: Creating new L1 chains (Core Protocol Dev).
+
+## Canonical Patterns (Best Practices)
+
+### 1. Smart Contracts (Solidity)
+
+- **Security First**: Use OpenZeppelin contracts (standard libraries).
+- **Checks-Effects-Interactions**: Update state _before_ external calls to prevent Reentrancy.
+- **Gas Optimization**: Use `calldata` vs `memory`, pack variables, avoid loops.
+
+### 2. Integration
+
+- **Wagmi/Viem**: Modern React hooks for wallet connection.
+- **Event Listening**: Index events for UI updates (The Graph or direct listeners).
+- **Error Handling**: Handle chain-specific errors (UserRejected, InsufficientFunds).
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Rolling Crypto**: Never invent your own encryption/hashing.
+- ❌ **Unchecked Calls**: Always check return values of low-level calls.
+- ❌ **Timestamp Dependence**: Block timestamps can be manipulated by miners.
+- ❌ **Private Logic**: Nothing on-chain is truly private.
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded Blockchain Pack."

package/.cursor/skills/_packs/common/pack-architecture.md

@@ -0,0 +1,40 @@
+# Pack: Architecture
+
+Tier: Core
+Load Cost: High
+Keywords: ddd, solid, patterns, planning, system-design
+
+## Scope
+
+This pack covers **Software Architecture and High-Level Planning**.
+
+- **Includes**: Domain-Driven Design (DDD), Modular Monolith, SOLID principles, Project Structure, Implementation Planning.
+- **Excludes**: Specific coding syntax.
+
+## Canonical Patterns (Best Practices)
+
+### 1. Structure (Modular Monolith)
+
+- **Feature Modules**: Group by feature (`features/billing/`), not type (`controllers/`).
+- **Encapsulation**: Modules expose a Public API (`index.ts`); internals are private.
+- **Dependency Rule**: Outer layers depend on inner layers (Core Domain).
+
+### 2. SOLID Principles
+
+- **SRP**: Single Responsibility per class/module.
+- **DIP**: Depend on abstractions (interfaces), not concretions.
+
+### 3. Planning
+
+- **Implementation Plan**: Always create `implementation_plan.md` before coding complex features.
+- **Task Breakdown**: Break tasks into < 1 hour chunks.
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Big Ball of Mud**: Circular dependencies between all modules.
+- ❌ **God Objects**: Classes that do everything.
+- ❌ **Coding without Plan**: Jumping into code for complex tasks.
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded Architecture Pack."

package/.cursor/skills/_packs/common/pack-devops.md

@@ -0,0 +1,43 @@
+# Pack: DevOps
+
+Tier: Domain
+Load Cost: Medium
+Keywords: docker, ci-cd, git, deployment, infrastructure, automation
+
+## Scope
+
+This pack covers **Development Operations, Infrastructure, and Automation**.
+
+- **Includes**: Git workflow, CI/CD pipelines (GitHub Actions), Docker containers, Deployment scripts, MCP Server management.
+- **Excludes**: Application code logic (Frontend/Backend).
+
+## Canonical Patterns (Best Practices)
+
+### 1. Git Workflow
+
+- **Conventional Commits**: `feat:`, `fix:`, `chore:`, `docs:`.
+- **Branching**: `main` (stable), `develop` (integration), `feature/xyz`.
+- **Clean History**: Rebase before merge preferred over merge commits for cleaner history.
+
+### 2. CI/CD (GitHub Actions)
+
+- **Fast Feedback**: Run lint and tests on every PR.
+- **Cache Dependencies**: Use `actions/cache` for npm/pnpm.
+- **Secret Management**: Use GitHub Secrets, never hardcode tokens in YAML.
+
+### 3. MCP (Model Context Protocol)
+
+- **Builder**: Use standard MCP SDKs.
+- **Management**: Register servers via config files, not ad-hoc commands.
+- **Security**: MCP servers run locally; trust bounds apply.
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Commit "WIP"**: Avoid vague messages. Squash WIP commits before merge.
+- ❌ **Broken Build**: Never merge code that fails CI.
+- ❌ **Manual Deploys**: Automate via pipeline whenever possible.
+- ❌ **Root Keys**: Never use root access keys in CI; use scoped service accounts.
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded DevOps Pack."

package/.cursor/skills/_packs/common/pack-productivity.md

@@ -0,0 +1,37 @@
+# Pack: Productivity
+
+Tier: Core
+Load Cost: Low
+Keywords: files, research, data, docs, organization
+
+## Scope
+
+This pack covers **General Productivity, File Management, and Data Handling**.
+
+- **Includes**: File searching, filtering, documentation organization, data analysis (CSV/JSON), learning patterns.
+
+## Canonical Patterns (Best Practices)
+
+### 1. File Management
+
+- **Search**: Use `scout` or `fd` to find files efficiently.
+- **Organization**: Keep directories flat where possible; group by domain.
+
+### 2. Documentation
+
+- **Proximity**: Keep docs close to code (`README.md` in folders).
+- **Updates**: Update docs _with_ code changes in the same PR.
+
+### 3. Data Analysis
+
+- **Validation**: Validate input data schema.
+- **Visualization**: Summarize data trends clearly.
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Stale Docs**: Documentation that contradicts code.
+- ❌ **Messy Workspace**: files scattered in root.
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded Productivity Pack."

package/.cursor/skills/_packs/common/pack-quality.md

@@ -0,0 +1,41 @@
+# Pack: Quality
+
+Tier: Domain
+Load Cost: Medium
+Keywords: testing, debug, review, security, performance, kaizen
+
+## Scope
+
+This pack covers **Quality Assurance, Testing, and Code Review**.
+
+- **Includes**: Unit Testing (Vitest/Jest), E2E (Playwright), Debugging strategies, Security audits, Performance profiling.
+- **Excludes**: Implementation details.
+
+## Canonical Patterns (Best Practices)
+
+### 1. Testing (TDD)
+
+- **Red-Green-Refactor**: Write test -> Fail -> Pass -> Clean.
+- **AAA Pattern**: Arrange, Act, Assert.
+- **Mocking**: Mock external boundaries (API, DB), test logic in isolation.
+
+### 2. Review
+
+- **Readability**: Code should be self-documenting.
+- **Maintainability**: Avoid complex abstractions early.
+- **Security**: Sanitize inputs, check authorization.
+
+### 3. Debugging
+
+- **Systematic Approach**: Reproduce -> Isolate -> Fix -> Verify.
+- **Logs**: Use structured logging.
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Testing Implementation**: Test behavior, not internal state.
+- ❌ **Flaky Tests**: Tests relying on timing/network without mocks.
+- ❌ **Premature Optimization**: Optimize only after profiling.
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded Quality Pack."

package/.cursor/skills/_packs/data/pack-ai.md

@@ -0,0 +1,41 @@
+# Pack: AI (Cognitive)
+
+Tier: Core
+Load Cost: Medium
+Keywords: llm, prompt, mcp, agents, reasoning, research
+
+## Scope
+
+This pack covers **AI Capabilities, Reasoning Strategies, and Research**.
+
+- **Includes**: Prompt Engineering, Agent Personas, Sequential Thinking, Research methodologies, Context Management.
+- **Excludes**: Specific implementation details (coding).
+
+## Canonical Patterns (Best Practices)
+
+### 1. Sequential Thinking
+
+- **Step-by-Step**: Break complex problems into atomic steps.
+- **Self-Correction**: Review previous steps before proceeding.
+- **Hypothesis Testing**: Formulate hypotheses and verifying them effectively.
+
+### 2. Research (Scout)
+
+- **Breadth-First**: Scan wide before diving deep.
+- **Source Verification**: Prefer official docs and code over generic articles.
+- **Synthesis**: Combine multiple sources into a coherent answer.
+
+### 3. Multi-Modal
+
+- **Images**: Analyze UI screenshots for design fidelity.
+- **Context**: Provide relevant context files, not whole folders.
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Hallucination**: Inventing APIs or libraries. Always verify.
+- ❌ **Looping**: Getting stuck in "I will fix" loops without changing strategy.
+- ❌ **Ignoring Constraints**: Forgetting project-specific rules (e.g., "No Tailwind").
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded AI Pack."

package/.cursor/skills/_packs/data/pack-data-science.md

@@ -0,0 +1,36 @@
+# Pack: Data Science
+
+Tier: Domain
+Load Cost: High
+Keywords: python, pandas, numpy, visualization, analysis, machine-learning, plot
+
+## Scope
+
+This pack covers **Deep Data Science and Analysis**.
+
+- **Includes**: Data manipulation (Pandas/Polars), Visualization (Matplotlib/Seaborn), Statistical Analysis, Basic ML (Scikit-learn).
+- **Primary Lang**: Python (assumed execution environment or script generation).
+
+## Canonical Patterns (Best Practices)
+
+### 1. Data Pipeline
+
+- **ETL**: Extract -> Transform -> Load. Clean data before analysis.
+- **Vectorization**: Use Pandas/NumPy vector operations, avoid Python loops.
+- **Reproducibility**: Use Jupyter Notebooks or scripts with seed fixing.
+
+### 2. Visualization
+
+- **Clarity**: Label axes, titles, and legends.
+- **Color**: Use accessible color palettes.
+- **Context**: Annotate outliers or key events.
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Mutating Data**: Modifying DataFrames in-place implicitly.
+- ❌ **Magic Numbers**: Hardcoding boundaries without explanation.
+- ❌ **Overfitting**: training on test data.
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded Data Science Pack."

package/.cursor/skills/_packs/mobile/pack-mobile.md

@@ -0,0 +1,40 @@
+# Pack: Mobile
+
+Tier: Domain
+Load Cost: Medium
+Keywords: react-native, expo, ios, android, mobile
+
+## Scope
+
+This pack covers **Mobile App Development** using React Native and Expo.
+
+- **Includes**: Native Components, Navigation (React Navigation/Expo Router), native modules, platform-specific code.
+- **Excludes**: Pure web components (DOM-based).
+
+## Canonical Patterns (Best Practices)
+
+### 1. Framework
+
+- **Expo Framework**: Preferred for new projects (File-based routing, Prebuild).
+- **NativeWind**: Tailwind CSS for React Native.
+
+### 2. Components
+
+- **Core Components**: Use `<View>`, `<Text>`, `<Pressable>` (not `<div>`, `<span>`).
+- **Lists**: Use `<FlatList>` or `<FlashList>` for long lists (virtualization).
+- **Images**: Use `expo-image` for caching and performance.
+
+### 3. Navigation
+
+- **Architecture**: Stack Navigator for flows, Tab Navigator for main shell.
+- **Deep Linking**: Configure schemes for external access.
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Web-only APIs**: Using `window`, `document`, or `localStorage` (Use `AsyncStorage`).
+- ❌ **Blocking JS Thread**: Heavy computations on UI thread. Use Worklets (Reanimated).
+- ❌ **Unoptimized Images**: Loading huge raw images.
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded Mobile Pack."

package/.cursor/skills/_packs/web/pack-backend.md

@@ -0,0 +1,61 @@
+# Pack: Backend
+
+Tier: Domain
+Load Cost: High
+Keywords: node, api, database, auth, payment, backend, server
+
+## Scope
+
+This pack covers **Backend Development** logic, API design, and Data Management.
+
+- **Includes**: API endpoints (Next.js App Router), Database interactions (Prisma/Drizzle), Authentication (Auth.js), Payments (Stripe), Business Logic.
+- **Excludes**: UI components, detailed DevOps/Infrastructure (see `pack-devops`).
+
+## Canonical Patterns (Best Practices)
+
+### 1. API Design
+
+- **Resource Oriented**: `GET /api/users`, `POST /api/users`.
+- **Validation**: Use Zod for all request body/param validation.
+- **Error Handling**: Return structured errors `{ code, message, details }`.
+- **Type Safety**: Share types between Backend and Frontend (Monorepo/RPC).
+
+### 2. Database Interactions
+
+- **ORM**: Use Prisma or Drizzle.
+- **Migrations**: Always verify migrations locally before commit.
+- **Indexing**: Add indexes for foreign keys and frequently searched columns.
+- **Connection Pooling**: Use connection pooling for Serverless environments.
+
+### 3. Authentication & Security
+
+- **Auth.js / Better-Auth**: Standard implementations preferred.
+- **Middleware**: Protect routes via middleware, not just client-side checks.
+- **RBAC**: Implement Role-Based Access Control logic in services.
+- **Secrets**: Never commit `.env`. Access via `process.env`.
+
+### 4. Payments (Stripe)
+
+- **Webhooks**: Always verify webhook signatures securely.
+- **Idempotency**: Handle duplicate webhook events gracefully.
+- **Client/Server Split**: Create PaymentIntent on server, confirm on client.
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Logic in Controllers**: Keep Route Handlers thin; move logic to Services/UseCases.
+- ❌ **N+1 Queries**: Loop queries inside loops. Use `include` or `Promise.all`.
+- ❌ **Trusting Client**: Never trust `req.body` without Zod validation.
+- ❌ **SQL Injection**: Never concat strings for queries. Use ORM or parametrized queries.
+- ❌ **Hardcoded IDs**: Do not rely on specific DB IDs in logic.
+
+## Specialized Modules
+
+### Shopify Integration
+
+- Use Shopify Admin API / Storefront API.
+- Respect rate limits.
+- Handle webhook HMAC verification strict.
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded Backend Pack."

package/.cursor/skills/_packs/web/pack-frontend.md

@@ -0,0 +1,66 @@
+# Pack: Frontend
+
+Tier: Domain
+Load Cost: High
+Keywords: react, nextjs, tailwind, ui, ux, components, frontend
+
+## Scope
+
+This pack covers all aspects of **Frontend Development** within the Anchi-Kit ecosystem.
+
+- **Includes**: Component architecture, UI styling (Tailwind), State Management, React Hooks, Next.js Routing, Client-Side logic, Accessibility (a11y), Animations.
+- **Excludes**: Backend API logic (see `pack-backend`), Database schemas, DevOps pipelines.
+
+## Canonical Patterns (Best Practices)
+
+### 1. Component Design
+
+- **Single Responsibility**: Each component should do one thing well.
+- **Composition over Inheritance**: Use children props and slots.
+- **Props Interface**: Always define explicit interfaces for props (TypeScript).
+- **Atomic Design**: Structure components as Atoms, Molecules, Organisms, Templates.
+
+### 2. Styling (Tailwind CSS)
+
+- **Utility-First**: Use Tailwind classes directly in markup.
+- **Tokens**: Use `theme('colors...')` for consistency.
+- **Responsive**: Mobile-first design (`min-width` default, `md:`, `lg:` overrides).
+- **ClassName Merging**: Use `cn()` utility (clsx + tailwind-merge) for conditional classes.
+
+### 3. State Management
+
+- **Local State**: `useState` for simple UI toggles.
+- **Server State**: Use `useQuery` or Next.js Server Components for data fetching.
+- **Global State**: Minimal use of Context/Zustand only when absolutely necessary (e.g., Theme, Auth User).
+
+### 4. Performance
+
+- **Image Optimization**: Use `next/image`.
+- **Lazy Loading**: Use `dynamic()` for heavy components below the fold.
+- **Memoization**: `useMemo`/`useCallback` only when profiling shows re-render issues.
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Hardcoded Styles**: Avoid `style={{ margin: 10 }}`. Use Tailwind classes.
+- ❌ **Prop Drilling**: Passing data through >3 layers. Use Composition or Context.
+- ❌ **Client-Side Heavy**: Making everything `"use client"`. Default to Server Components.
+- ❌ **Magic Numbers**: Using arbitrary pixel values (e.g., `width: 317px`). Use design tokens.
+- ❌ **Huge Components**: Files > 300 lines should likely be split.
+
+## Specialized Modules
+
+### 3D / WebGL (Three.js)
+
+- Use `@react-three/fiber`.
+- Keep canvas logic separate from DOM UI.
+- optimize models (GLB/GLTF) via `gltfjsx`.
+
+### Mobile (React Native / Expo)
+
+- Use `StyleSheet.create` or strict Tailwind integration (NativeWind).
+- Handle platform differences (`Platform.OS`).
+- Avoid web-only APIs is critical paths.
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded Frontend Pack."

package/.cursor/skills/_packs/web3/pack-blockchain.md

@@ -0,0 +1,37 @@
+# Pack: Blockchain
+
+Tier: Domain
+Load Cost: High
+Keywords: blockchain, web3, solidity, smart-contract, ethereum, crypto, wallet
+
+## Scope
+
+This pack covers **Web3 and Blockchain Development**.
+
+- **Includes**: Smart Contract development (Solidity), Interaction (Ethers.js/Viem), Testing (Hardhat/Foundry), Security (Reentrancy prevention).
+- **Excludes**: Creating new L1 chains (Core Protocol Dev).
+
+## Canonical Patterns (Best Practices)
+
+### 1. Smart Contracts (Solidity)
+
+- **Security First**: Use OpenZeppelin contracts (standard libraries).
+- **Checks-Effects-Interactions**: Update state _before_ external calls to prevent Reentrancy.
+- **Gas Optimization**: Use `calldata` vs `memory`, pack variables, avoid loops.
+
+### 2. Integration
+
+- **Wagmi/Viem**: Modern React hooks for wallet connection.
+- **Event Listening**: Index events for UI updates (The Graph or direct listeners).
+- **Error Handling**: Handle chain-specific errors (UserRejected, InsufficientFunds).
+
+## Anti-Patterns (What to Avoid)
+
+- ❌ **Rolling Crypto**: Never invent your own encryption/hashing.
+- ❌ **Unchecked Calls**: Always check return values of low-level calls.
+- ❌ **Timestamp Dependence**: Block timestamps can be manipulated by miners.
+- ❌ **Private Logic**: Nothing on-chain is truly private.
+
+## Load Confirmation
+
+> Rule: If you use this pack, you must explicitly state: "Loaded Blockchain Pack."

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "anchi-kit",
-  "version": "1.2.2",
+  "version": "1.2.3",
   "description": "The Ultimate AI-Native Toolkit for Cursor & Gemini. Installs into any existing project. Includes commands, agents, skills, and architecture presets.",
   "main": "src/cli.js",
   "bin": {
@@ -41,6 +41,7 @@
     "scripts/",
     ".cursor/commands/",
     ".cursor/agents/",
+    ".cursor/skills/_packs/",
     ".cursor/skills/*/SKILL.md",
     ".cursor/skills/*/references/",
     ".cursor/workflows/",
@@ -55,6 +56,7 @@
     ".cursor/UX_GUIDELINES.md",
     ".antigravity/commands/",
     ".antigravity/agents/",
+    ".antigravity/skills/_packs/",
     ".antigravity/skills/*/SKILL.md",
     ".antigravity/skills/*/references/",
     ".antigravity/workflows/",