LLM & Agents

This skill should be used when the user asks about "nativephp api", "camera api", "device api", "biometrics", "geolocation", "scanner api", "microphone api", "nativephp_call", "god method", "bridge function", "SecureStorage", "Dialog api", "Share api", "PushNotifications api", "Network status", "Browser api", "Haptics", "File api", or needs to use any NativePHP native functionality in their app.

エージェント、メモリ、ツール統合パターンを備えたLangChainフレームワークを使用してLLMアプリケーションを設計します。LangChainアプリケーションの構築、AIエージェントの実装、または複雑なLLMワークフローの作成時に使用します。

Expert guidance for vibe-validate, an LLM-optimized validation orchestration tool. Use when working with vibe-validate commands, configuration, pre-commit workflows, or validation orchestration in TypeScript projects.

Create and maintain CLAUDE.md files for Claude Code projects. Use when users want to create a new CLAUDE.md, improve an existing one, audit CLAUDE.md quality, set up Claude Code configuration for a repository, or optimize agent context for monorepos. Triggers on requests like "create CLAUDE.md", "improve my CLAUDE.md", "set up Claude Code", "optimize agent context", or "audit my project documentation for Claude".

Expert skill for prompt engineering and task routing/orchestration. Covers secure prompt construction, injection prevention, multi-step task orchestration, and LLM output validation for JARVIS AI assistant.

Expert SurrealDB developer specializing in multi-model database design, graph relations, document storage, SurrealQL queries, row-level security, and real-time subscriptions. Use when building SurrealDB applications, designing graph schemas, implementing secure data access patterns, or optimizing query performance.

Execute multiple independent tasks simultaneously using parallel agent coordination to maximize throughput and minimize execution time. Use when tasks have no dependencies, results can be aggregated, and agents are available for concurrent work.

Expert in batch cooking, meal planning, food storage, and efficient kitchen workflows

Configure Claude Code sandbox network isolation with trusted domains, custom access policies, and environment variables

Create and manage Dexie/IndexedDB repositories with type-safe interfaces, converters, and standardized CRUD operations. Use when (1) adding entity storage, (2) implementing save/load/delete operations, (3) designing database schema and indexes, (4) converting between database (Db*) and domain types, (5) handling database errors or migrations, (6) using existing repositories (SettingsRepository, WorkoutsRepository, TemplatesRepository, CustomExercisesRepository, BenchmarksRepository, ActiveWorkoutRepository). Triggers include "database", "repository", "save data", "fetch from database", "delete from storage", "database schema", "database table", "indexes", "migration", "persist", "convert workout", "converter", "buildPartialUpdate", "mock repository", "database error", "bulk operations", "import/export", or specific repository names.

Setup and configure React Grab to let you click elements in your running app and send them directly to Claude Code for editing. Use when setting up React Grab, when the user wants to click UI elements to modify them, or mentions wanting visual element selection for development. Makes AI coding agents 66% faster.

Scaffolds complete Claude Code plugin structures with all necessary directories, manifest, and documentation. Activates when user wants to create a new plugin, package features together, or prepare for marketplace distribution. Creates plugin.json, directory structure, and starter documentation. Use when user mentions "create plugin", "new plugin", "scaffold plugin", "plugin structure", or "share via marketplace".

Dynamic prompt registry with unified categorical syntax. Supports @skills:discover(), @skills:compose(A⊗B), Reader monad for runtime lookup, and quality-tracked prompt libraries. Use for meta-prompts referencing sub-prompts, building prompt libraries, implementing deferred resolution, or composing templates dynamically.

Decentralized database protocol with zero-knowledge proofs and permanent storage on Arweave

Create effective Agent Skills following the agentskills.io specification. Use when creating new skills, reviewing existing skills, or improving skill descriptions. Covers SKILL.md format, naming, descriptions, structure, progressive disclosure, and testing patterns.

Generic desktop notification skill for agents. Send desktop notifications when tasks are complete (or when user input/errors block progress). By default, assume that all jobs will require a notification unless the user says otherwise.

Interactive Flutter Widget implementation assistant that guides you through widget implementation decisions by asking structured questions about state management, widget type (screen/component), and state sharing. Use this skill when you need help determining the appropriate Flutter widget architecture for your implementation.

Optimize system prompts for Claude Code agents using proven prompt engineering patterns. Use when users request prompt improvement, optimization, or refinement for agent workflows, tool instructions, or system behaviors.

Run the case feedback script to submit feedback for the current conversation. Only use when invoked by case-feedback-agent.

3D HUD rendering with Three.js and TresJS for JARVIS AI Assistant