AI & ML

majiayu000 Bundle

Integrate Gemini API with correct current SDK (@google/genai v1.27+, NOT deprecated @google/generative-ai). Supports text generation, multimodal (images/video/audio/PDFs), function calling, and thinking mode. 1M input tokens. Use when: integrating Gemini API, implementing multimodal AI, using thinking mode for reasoning, function calling with parallel execution, streaming responses, deploying to Cloudflare Workers, building chat, or troubleshooting SDK deprecation, context window, model not found, function calling, or multimodal format errors. Keywords: gemini api, @google/genai, gemini-2.5-pro, gemini-2.5-flash, gemini-2.5-flash-lite, gemini-3-pro-preview, multimodal gemini, thinking mode, google ai, genai sdk, function calling gemini, streaming gemini, gemini vision, gemini video, gemini audio, gemini pdf, system instructions, multi-turn chat, DEPRECATED @google/generative-ai, gemini context window, gemini models 2025, gemini 1m tokens, gemini tool use, parallel function calling, compositional function call

majiayu000 Bundle

dbt project structure using medallion architecture (bronze/silver/gold layers). Use this skill when planning project organization, establishing folder structure, defining naming conventions, implementing layer-based configuration, or ensuring proper model dependencies and architectural patterns.

majiayu000 Bundle

Writing SKILL.md, .claude/commands/*.md, Task tool prompts, systemPrompt variables, prompt templates, few-shot examples, agent instructions. Content where the reader is an LLM.

majiayu000 Bundle

Reviews BubbleTea TUI code for proper Elm architecture, model/update/view patterns, and Lipgloss styling. Use when reviewing terminal UI code using charmbracelet/bubbletea.

majiayu000 Bundle

Structure LLM context using XML tags for maximum attention efficiency. Use when writing system prompts, organizing complex instructions, or engineering context for Claude agents. Produces semantically meaningful XML that creates clear attention anchors.

majiayu000 Bundle

Build stateful chatbots with OpenAI Assistants API v2 - Code Interpreter, File Search (10k files), Function Calling. ⚠️ Deprecated (sunset August 26, 2026); use openai-responses for new projects. Use when: maintaining legacy chatbots, implementing RAG with vector stores, or troubleshooting "thread has active run", vector store delays, polling timeouts, or file upload errors.

majiayu000 Bundle

Multi-agent orchestration system that coordinates specialized agents (PM, Architect, DevOps, QA, Tech Lead, Security) to work together on complex tasks. Implements hierarchical orchestrator-worker pat

majiayu000 Bundle

SonarQube integration via MCP for Java projects. The MCP server runs remotely and requires no local setup. Use natural language to query issues, analyze code, and check quality gates.

majiayu000 Bundle

Leverage Perplexity CLI for web research, deep reasoning, and independent validation. Use when you need current/external knowledge, want a second opinion before decisions, or want to delegate complex analytical thinking to specialized reasoning models. Complements (not replaces) standard web tools. Can be invoked directly or spawned as a subagent.

majiayu000 Bundle

🗣️ Speech — Text-to-Speech & Speech Recognition

majiayu000 Bundle

Search the web and fetch page content.

majiayu000 Bundle

Send and receive Signal messages via MCP. Use this skill when you need to interact with Signal messenger - sending messages, reading conversations, or automating Signal-based workflows.

majiayu000 Bundle

- Run at the start of every session and whenever prompt sizes or tool calls begin to spike.

majiayu000 Bundle

Fetch and summarize authoritative external sources into actionable guidance using MCP helpers. EXCLUSIVE to researcher agent.

majiayu000 Bundle

Erstellt wöchentliche Statusanfrage-E-Mail-Entwürfe an externe Bevollmächtigte (Sozietät/Korrespondenzanwalt) für alle aktiven Prozessmandate im Portfolio — Markdown-Entwurf je Mandat, optional Outlook-Entwurf bei verfügbarer MCP-Integration. Lädt, wenn der Nutzer Statusanfragen an externe Bevollmächtigte erstellen, wöchentliche Rückfragen bei der Außensozietät generieren oder mandatsbezogene Anfragen-E-Mails aus dem Portfolio-Protokoll heraus entwerfen möchte.

majiayu000 Bundle

Design hook-based event systems for ADW observability. Use when implementing real-time event broadcasting, creating hook pipelines, or building agent activity monitoring.

majiayu000 Bundle

Central authority for Claude Code permissions. Covers permission system (tiered model, read-only, bash, file modification), permission rules (allow, ask, deny), rule precedence, permission modes (default, acceptEdits, plan, bypassPermissions), tool-specific permission rules (Bash patterns, Read/Edit patterns, WebFetch domains, MCP tools), /permissions command, and additionalDirectories. Assists with configuring permission rules, understanding permission modes, and troubleshooting permission issues. Delegates 100% to docs-management skill for official documentation.

majiayu000 Bundle

LLM cost tracking with Langfuse for cached responses. Use when monitoring cache effectiveness, tracking cost savings, or attributing costs to agents in multi-agent systems.

majiayu000 Bundle

CRITICAL: Before deploying any skill, spawn a reviewer agent with REVIEW.md to audit allowed-tools. Overly permissive tool access (e.g., Bash(git:)) can cause data loss or leak secrets.

majiayu000 Bundle

Prompt Engineering SKILL

majiayu000 Bundle

Run Codex CLI, Claude Code, OpenCode, or Pi Coding Agent via background process for programmatic control.

majiayu000 Bundle

This document describes the GitHub custom agent file format used by GitHub Copilot to provide repository-specific or task-specific instructions to AI coding agents.

majiayu000 Bundle

Use CodexBar CLI local cost usage to summarize per-model usage for Codex or Claude, including the current (most recent) model or a full model breakdown. Trigger when asked for model-level usage/cost data from codexbar, or when you need a scriptable per-model summary from codexbar cost JSON.

majiayu000 Bundle

- Quick Start

majiayu000 Bundle

Compile or lint a persistent Mayor-style goal that ratchets a bead graph through bounded RPI experiments toward one larger outcome. Triggers: "craft a goal prompt", "turn this into a goal", "create a bounded goal", "lint this goal", "is this goal safe".

majiayu000 Bundle

Use when user asks "what do you know about X", when planning complex work that spans multiple topics, when investigating how concepts connect across projects, or when simple memory queries don't provide enough context. Deep traversal of Forgetful MCP knowledge graph (mcp__forgetful__* tools).

majiayu000 Bundle

Manage semantic versioning, changelog generation, and release notes. Auto-generates entries from completed issues or git diff.

majiayu000 Bundle

Use when coordinating multiple AI agents in real-time - provides inter-agent messaging via Rust PTY wrapper with file-based protocol and reliability features

majiayu000 Bundle

LangGraph supervisor-worker pattern. Use when building central coordinator agents that route to specialized workers, implementing round-robin or priority-based agent dispatch.

majiayu000 Bundle

LangGraph human-in-the-loop patterns. Use when implementing approval workflows, manual review gates, user feedback integration, or interactive agent supervision.

majiayu000 Bundle

Manage coordinator daemon tasks, approve/reject work, monitor autonomous agents. Use when user asks to delegate tasks, check task status, review agent work, manage the coordinator, or use GitHub-driven approval workflow.

majiayu000 Bundle

コマンドのセキュリティ設計（allowed-tools制限/disable-model-invocation/機密情報保護）を整理し、安全な実行フローと権限制御を支援するスキル。 セキュリティ要件、検証手順、テンプレート運用を一貫して整理する。 Anchors: • Web Application Security (Andrew Hoffman) / 適用: 脅威モデリング / 目的: セキュア設計指針 • OWASP Top 10 / 適用: 一般的脅威の整理 / 目的: リスク評価基準 Trigger: Use when designing secure command execution, restricting tools, or preventing unsafe automated operations. command security, allowed-tools, disable-model-invocation, secret protection

majiayu000 Bundle

This skill provides complete coverage of Google Gemini embeddings API (gemini-embedding-001) for building RAG systems, semantic search, document clustering, and similarity matching. Use when implementing vector search with Google's embedding models, integrating with Cloudflare Vectorize, or building retrieval-augmented generation systems. Covers SDK usage (@google/genai), fetch-based Workers implementation, batch processing, 8 task types (RETRIEVAL_QUERY, RETRIEVAL_DOCUMENT, SEMANTIC_SIMILARITY, etc.), dimension optimization (128-3072), and cosine similarity calculations. Prevents 8+ embedding-specific errors including dimension mismatches, incorrect task types, rate limiting issues (100 RPM free tier), vector normalization mistakes, text truncation (2,048 token limit), and model version confusion. Includes production-ready RAG patterns with Cloudflare Vectorize integration, chunking strategies, and caching patterns. Token savings: ~60%. Production tested. Keywords: gemini embeddings, gemini-embedding-001, go

majiayu000 Bundle

Add OAuth authentication to MCP servers on Cloudflare Workers. Uses @cloudflare/workers-oauth-provider with Google OAuth for Claude.ai-compatible authentication. Use when building MCP servers that need user authentication, implementing Dynamic Client Registration (DCR) for Claude.ai, or replacing static auth tokens with OAuth flows. Prevents CSRF vulnerabilities, state validation errors, and OAuth misconfiguration.

majiayu000 Bundle

Reference guide for using PROXY_MODE with external AI models. Use when running multi-model reviews, understanding which agents support PROXY_MODE, or debugging external model integration issues.

majiayu000 Bundle

⚡ PRIMARY TOOL for: 'comprehensive audit', 'deep analysis', 'full codebase review', 'multi-perspective investigation', 'complex questions'. Combines ALL detective perspectives (architect+developer+tester+debugger). Uses Opus model. REPLACES grep/glob entirely. Uses claudemem v0.3.0 AST with ALL commands (map, symbol, callers, callees, context). GREP/FIND/GLOB ARE FORBIDDEN.