All prompts

Design a comprehensive stress testing plan for an AI Agent's memory system, evaluating retrieval accuracy, forgetting curves, context window utilization, and cross-session consistency.

Instruct the AI Agent to automatically loop through and execute all items in the PRD until completion. Suitable for autonomous development scenarios with coding agents like Claude Code or Codex.

Simulate a hedge fund team composed of multiple AI Agents, with roles such as analysts, risk managers, and traders making investment decisions.

Comprehensively evaluate security isolation solutions for AI Agent code execution sandboxes, covering technology selections such as containers, microVMs, and WASM.

Compare mainstream Go-based LLM development frameworks (such as eino, langchaingo, etc.) and output technology selection recommendations and project architecture designs.

Simulate a senior SRE engineer to systematically analyze the root causes of production incidents and generate a structured RCA report.

Product requirements, technical architecture, and deployment plan for building an open-source AI voice synthesis studio from scratch, supporting voice cloning and style transfer.

Deeply understand the principles of Speculative Decoding technology and design the optimal acceleration solution for LLM inference services.

Input a note topic to automatically generate 10 Xiaohongshu-style viral titles and cover copies, including high click-through rate elements like emojis, numbers, and pain-point keywords.

Automatically generate standard operating procedures (Runbooks) for incident response based on your system architecture and alert information, including diagnostic steps, fix commands, and escalation paths.

Input agent execution logs or error messages to automatically analyze workflow bottlenecks, reasons for tool call failures, and provide fix suggestions and optimization plans. Suitable for frameworks like LangChain, CrewAI, and AutoGen.

Transform any topic, article, or idea into a Xiaohongshu-style image-text note copy, including title, body, tags, and cover suggestions.

Input MCP Server tool definitions to automatically generate natural language test cases covering normal, edge, and exception scenarios, helping to quickly verify the quality of agent tool calls.

Compare and evaluate the performance, quality, and deployment costs of different Speculative Decoding solutions, generating a selection recommendation report.

Generates a complete CLAUDE.md configuration file based on your project needs, integrating best practices for LLM coding pitfalls summarized by Andrej Karpathy, ensuring Claude Code outputs more stable and reliable code.

Input your hardware configuration and usage scenario, and the AI will generate a complete local large model deployment decision tree, recommending the most suitable model, quantization scheme, and inference framework.

Generates a context window optimization checklist for AI coding agents, reducing token consumption and improving code generation quality

Systematic comparative analysis of open-source AI clients generates technology selection decision reports

Used to evaluate AI Agent conversation quality, scoring from multiple dimensions such as accuracy, relevance, security, and formatting standards, generating structured evaluation reports suitable for quality control before Agent products go live.

Input a complex task goal, let AI automatically break it down into executable sub-tasks, generate an execution plan with dependencies and structured log templates, suitable for Agent development and debugging.

Transform any topic or article into structured knowledge cards containing core concepts, key data, analogy explanations, and memory anchors, suitable for learning and content creation.

Input any GitHub open source project information to automatically generate a 5-minute quick start guide for beginners

A Manus-style Markdown persistent planning approach helps AI Agents break down complex tasks step by step and track execution progress

Simulate a professional Bloomberg Terminal style, using AI to analyze stocks, futures, and macroeconomic data, generating professional-grade investment research reports and visualization chart recommendations.

Directly convert a single-sentence natural language requirement into a complete MCP Server interface design document, including tool definitions, parameter schemas, error handling, and example calls.

Input natural language requirements to automatically generate an MCP protocol-compliant Server interface design document, including tool definitions, parameter schemas, and example calls.

Describe your business scenario in natural language to automatically generate multi-model routing strategy configurations, achieving the optimal balance between cost and quality.

Input a topic to automatically generate a complete workflow prompt for short video production, including script, storyboard, voiceover copy, subtitles, and publishing plan.

Automatically generate deployment plans for multi-model hybrid inference clusters based on business scenarios and hardware conditions, including model allocation, routing strategies, and cost estimates.

Helps you systematically analyze and optimize the context window usage efficiency of AI coding agents (such as Claude Code, Cursor, Copilot), generating specific slimming plans and configuration suggestions.

Batch converts multimodal documents such as PDFs, images, and web pages into structured Markdown knowledge bases, supporting direct integration with RAG systems.

Evaluates the maturity of an AI Agent's self-evolution capabilities and provides evolution asset design proposals based on the GEP (Gene Expression Programming) framework, including implementation suggestions for a three-layer architecture of Genes, Capsules, and Events.

Addresses the context window explosion problem for AI coding agents (such as Claude Code, Cursor, Codex) by generating sandboxing strategies for tool outputs and token recycling plans to achieve 90%+ context savings.

A complete solution for building a local voice cloning studio based on open-source solutions (Voicebox/Kokoro/Chatterbox, etc.), including multi-TTS engine comparison evaluation, voice cloning quality assessment, and privacy compliance checks.

Input ML papers, extract core methods and innovations, and generate experimental reproduction plans

One piece of copy, rewritten across multiple platforms, adapted for Xiaohongshu, Twitter, LinkedIn, and official accounts

Input a list of pending tasks to the AI, which will automatically sort them based on urgency, dependencies, and estimated duration, outputting an execution plan with time suggestions.

Decompose complex project requirements into multiple Agent subtasks, assign them to different platforms (Web/Desktop/Terminal) for coordinated execution, and generate a Kanban board and dependency graph.

Paste any open-source project's README to receive an attractiveness score and a professional-grade rewritten version, significantly boosting Star conversion rates.

Evaluate the first-screen effect of your GitHub README using the 3-second rule, receive an attractiveness score, and get an automatic rewrite to help open-source projects boost Star conversion rates.

For Mixture of Experts (MoE) architecture models, generates detailed inference deployment plans, including hardware selection, communication optimization, and cost estimation.

Generate a comprehensive performance benchmark plan for LLM inference services (such as vLLM, SGLang, TensorRT-LLM), including test metrics, load models, and result analysis templates.

Based on historical project data, community activity, and external events, predict the Star growth trend of open-source projects, generating community operation growth models and strategic recommendations.

Leave your work logbook to AI, which automatically distills it into a clear, highlighted-point weekly report suitable for your boss or team.

Generates comprehensive desktop operation test scenarios for Computer-Use Agents (CUAs), including multi-step UI interactions, exception handling, and evaluation metrics, suitable for macOS/Linux/Windows sandbox environments.

Uses LLMs to extract structured entities from unstructured text, supporting custom schemas and few-shot examples, suitable for document parsing, report analysis, and other scenarios.

Input an initial prompt to automatically analyze its weaknesses and generate 3 optimized variants, complete with improvement rationale and expected effect comparisons, to help rapidly iterate on prompt quality.

Describe your microservice deployment needs in natural language; the AI automatically generates a complete Helm Chart, including values.yaml, templates, and NOTES.txt.