All prompts

Systematically diagnose performance bottlenecks, error loops, and token waste in AI Agent workflows, outputting executable optimization plans.

Assess the feasibility of deploying large models on edge devices for specific application scenarios, including model selection, quantization strategies, hardware requirements, and performance estimates.

Helps you write standardized SKILL.md files for AI coding assistants (Claude Code, Codex CLI, etc.), including trigger conditions, tool invocation specs, and examples.

Give AI the key file contents of a GitHub repository to quickly generate an architecture understanding document and Mermaid diagram.

Transform a vague product idea into a complete technical solution, including tech stack selection, file structure, and core code skeleton, ready to be handed off to an AI coding assistant.

Conduct competitive analysis for a specified open-source project and generate differentiated positioning recommendations and growth strategies.

Design an end-to-end automated workflow for translating comics/image-text content, combining OCR, machine translation, and image editing.

Evaluates the quality of translated and localized content, checking for cultural adaptation, terminology consistency, and naturalness, then outputs a detailed review report.

Designs fingerprint isolation solutions for multiple browser profiles, covering best practices for proxies, cookies, extensions, and identity management.

Quickly evaluate the technical value, community activity, security risks, and applicable scenarios of a GitHub open-source project, generating a structured one-page report.

Helps you systematically track and analyze token consumption and costs across multiple AI coding assistants (Claude Code, Codex, Cursor, etc.), generating optimization suggestions.

Design AI Agents as independently deployable and callable microservices, including routing, coordination, memory management, and audit trails, suitable for enterprise-grade AI backend architecture design.

Design a dynamic memory system based on knowledge graphs for AI Agents, supporting temporal awareness, fact tracking, and incremental updates.

OCR recognition and structured data output for complex tables, handwriting, and mixed-layout documents

Help developers debug Agent behavior in multi-tool coordination scenarios, including systematic diagnosis of tool selection, parameter passing, and error recovery.

Convert any GitHub repository into a Mermaid knowledge graph to quickly understand code architecture, module dependencies, and core logic.

Simulate multiple AI expert roles to conduct a comprehensive inspection of a code repository, covering security, performance, architecture, documentation, and other dimensions, generating a structured inspection report.

Describe requirements in natural language to automatically generate a complete multi-agent workflow, including tool creation, agent definition, and workflow orchestration, without writing any code.

Evaluate different AI Agent sandbox execution environments (Flue, E2B, Modal, Firecracker, etc.), output selection comparison matrices, and production-level security configuration schemes.

A prompt specifically designed for testing and debugging Model Context Protocol (MCP) Servers, covering tool discovery, invocation validation, and troubleshooting.

Input your AI application scenario to automatically estimate Token usage and API costs, and provide multi-model routing optimization suggestions.

Quickly assess the technical value, community health, and practicality of a GitHub open-source project, outputting a structured evaluation report.

Input research hypotheses and constraints to automatically generate a complete experimental design plan, including variable control, data collection plans, and statistical analysis methods.

Use a structured method to design high-quality synthetic datasets, including field definitions, distribution control, dependency relationships, and quality verification.

Generate an MCP Server integration solution based on Google Agent Development Kit (ADK) for your AI Agent project, including code templates and deployment guides.

Analyze the proportion, quality, and risk of AI-generated code in a repository to generate a structured audit report.

Design a cost monitoring solution for LLM API calls, including token usage tracking, model routing optimization, and budget alerts.

Rapidly generate MCP database tool server configurations for your AI Agent, supporting PostgreSQL, MySQL, SQLite, and more, with one-click connection to MCP clients like Claude Code and Gemini CLI.

Input your AI coding assistant usage data to generate a detailed token consumption analysis report, including model comparisons, cost trends, and optimization suggestions.

Design character personas and dialogue styles for voice AI applications, supporting roleplay and emotional expression control in real-time voice interaction scenarios.

Design anti-detection solutions for AI Agent network data collection scenarios, evaluate technology combinations such as stealth browsers, fingerprint disguises, and agent pools, and generate practical crawler stealth architecture solutions.

Design AI-assisted multi-platform content monetization strategies for content creators, including topic planning, platform adaptation, traffic conversion paths, and monetization model design.

Design a locally deployed document semantic search system for individuals or teams, covering embedding model selection, vector databases, and retrieval strategies.

Input your Agent tool list and common error scenarios to automatically generate a complete error handling decision tree, including retry strategies, fallback plans, and user notification templates.

Plan GPU configuration, batching strategies, and cost optimization for LLM inference clusters based on model specs and traffic demands.

Helps you analyze performance bottlenecks in Rust-based ML inference services, offering targeted suggestions such as memory layout optimization, SIMD vectorization, and asynchronous batch processing

Quickly generate product plans for voice AI applications based on business scenarios, including technology selection, architecture design, and cost estimation.

Design a time-series knowledge graph-based memory system for AI Agents, supporting fact version tracking, entity evolution, and hybrid retrieval.

Help developers evaluate and select suitable AI Agent gateway/proxy layer solutions, covering key dimensions such as routing, security, and observability.

Go beyond line-level diffs to analyze the scope and risk of code changes at the semantic entity level (functions, classes, methods, etc.).

Based on Andrej Karpathy's observations on LLM coding pitfalls, generate a set of CLAUDE.md / AGENTS.md coding rules to improve the code quality of AI coding assistants.

Convert natural language editing intents into structured video editing instructions, suitable for AI-driven video editing workflows.

Design a rigorous A/B test experiment plan for your LLM application, including sample size calculation, evaluation metric definition, statistical significance testing, and result visualization code.

For visual AI or NLP datasets, systematically evaluate annotation quality, distribution bias, duplicate samples, and other issues, and generate actionable cleaning solutions

Based on the user's hardware configuration (GPU/CPU/memory), it automatically generates the optimal local LLM inference service deployment script, supporting llama.cpp, vLLM, Ollama solution selection and parameter optimization.

Design plugin/extension solutions for AI assistant platforms (OpenAI GPTs, Claude MCP, Gemini Extensions, etc.), and output complete interface definitions, authentication plans, and deployment plans.

Analyze the structure of system prompts from mainstream AI models, learn their design patterns, and help you write higher-quality system prompts.

Convert natural language descriptions into precise AI video editing instructions, supporting scene decomposition, transition design, and style consistency control.