Submit a tool
Skills12 AI Model Guides

AI Prompt Guides

Master how to communicate effectively with modern AI systems using battle-tested prompt architectures and copyable templates.

Dedicated AI Platform Guides

Tailored prompt structures that leverage the unique architecture, context windows, and tools of each model.

ChatGPT

Prompt Guide & Architecture

Master conversational reasoning, Custom Instructions, and the OpenAI structured prompting pattern.

Read Full Guide

Claude

Prompt Guide & Architecture

Harness Claude's massive context window, XML tag structuring, and Artifact rendering precision.

Read Full Guide

Cursor

Prompt Guide & Architecture

Maximize codebase indexing, @Files context referencing, and targeted diff generation.

Read Full Guide

DeepSeek-R1 & V3

Prompt Guide & Architecture

Maximize open-weights reasoning power, chain-of-thought deliberation, and mathematical precision.

Read Full Guide

Flux.1 & Black Forest Labs

Prompt Guide & Architecture

Master 12B-parameter rectified flow transformers, photorealistic prompt engineering, and typography generation.

Read Full Guide

Gemini

Prompt Guide & Architecture

Leverage 2M+ token multimodal reasoning, Google Workspace grounding, and native YouTube/Search integration.

Read Full Guide

GitHub Copilot

Prompt Guide & Architecture

Master contextual repo-wide prompting, inline completions, slash commands, and Copilot Edits.

Read Full Guide

Microsoft 365 Copilot

Prompt Guide & Architecture

Drive enterprise automation across Excel data modeling, Word executive syntheses, and Teams meeting summaries.

Read Full Guide

Midjourney

Prompt Guide & Architecture

Master aesthetic prompt parameters, lighting tokens, composition tags, and stylize flags.

Read Full Guide

Mistral AI & Le Chat

Prompt Guide & Architecture

Leverage European frontier open weights, Codestral code generation, and concise multilingual precision.

Read Full Guide

Perplexity

Prompt Guide & Architecture

Unlock precision research, real-time web grounding, academic source filtering, and cited answers.

Read Full Guide

v0 by Vercel

Prompt Guide & Architecture

Generate clean React Server Components, Tailwind CSS styling, and shadcn/ui component interfaces.

Read Full Guide
The Universal Prompt Engineering Architecture

The 7 Core Components of High-Performance Prompts

Rather than relying on vague "secret magic words," state-of-the-art models perform best when instructions follow an explicit, structured cognitive scaffold:

ROLE

1. Role & Persona

Define who the AI is acting as (e.g. 'Principal Cloud Architect', 'Senior Investigative Journalist'). Sets tone and technical depth.

CONTEXT

2. Context & Environment

Provide codebase specs, stack versions, user constraints, and background business circumstances.

OBJECTIVE

3. Clear Objective

State the single concrete goal without ambiguity. Avoid open-ended, unfocused tasks.

CONSTRAINTS

4. Negative Constraints

Explicitly declare boundaries ('Do not use external dependencies', 'Keep response under 300 words').

OUTPUT FORMAT

5. Output Format

Specify exact structure: Markdown table, strict JSON schema, or numbered implementation steps.

EXAMPLES

6. Few-Shot Examples

Supply 1-2 concrete pairs of sample inputs and desired outputs when exact style matching is required.

ITERATION

7. Follow-Up Iteration

Refine responses sequentially rather than restarting: 'Critique your code against performance edge cases, then patch it.'

Proven Prompt Templates

Ready-to-copy, battle-tested prompt templates across practical engineering and business use cases.

AllCodingResearchWritingStudySEOMarketingBusinessDesignImage GenerationData AnalysisProductivity
CodingMistral AI & Le Chat

Codestral Fill-in-the-Middle Code Completion

Executes Fill-in-the-Middle (FIM) completions matching Codestral's native training architecture.

Prefix: [INSERT PRECEDING CODE] Suffix: [INSERT SUCCEEDING CODE] Task: Generate only the middle implementation to connect prefix and suffix cleanly without regressions.
Ready to paste
CodingChatGPT

Full-Stack Bug Diagnosis & Targeted Fix

Diagnose and fix complex full-stack bugs with strict minimal diff constraints.

You are a Principal Full-Stack Engineer specializing in Next.js (App Router), TypeScript, and SQLite. Context: I am observing an unexpected behavior in my application. Below is the relevant code and error log: [INSERT CODE & ERROR LOG] Task: Diagnose the root cause of this issue, explain why it happened, and provide the smallest safe code fix. Constraints: - Do not rewrite unaffected functions or components. - Maintain 100% type safety without using 'any'. - Preserve existing comments and architecture. Output Format: 1. Root Cause Summary (2-3 sentences) 2. Minimal Replacement Diff 3. Regression Protection Checklist
Ready to paste
CodingDeepSeek-R1 & V3

High-Performance Algorithm Optimization & Benchmark

Rewrites performance-critical algorithms for maximum cache locality and zero-allocation execution.

You are a Low-Level Systems Performance Engineer. Analyze the following algorithmic hot path: [INSERT CODE] Task: Optimize the routine to minimize CPU cache misses, eliminate heap allocations on the hot path, and improve throughput. Provide: - Asymptotic complexity comparison - Optimized implementation - Detailed rationale for cache line alignment and branch predictability.
Ready to paste
CodingClaude

Large Codebase Architecture Refactor Blueprint

Produces thorough, phased architectural refactoring plans with XML structure.

<role>You are a Principal Software Architect.</role> <context> [INSERT SYSTEM OVERVIEW & CODE] </context> <instructions> Analyze the current implementation for maintainability, coupling, and scalability bottlenecks. Produce a phased, safe refactoring roadmap. </instructions> <constraints> - Maintain backwards compatibility. - Ensure every phase has an independent verification test suite. </constraints> <output_format> Structure your response in clear XML tags: <analysis>Current architectural evaluation</analysis> <phases>Step-by-step phased execution plan</phases> <safeguards>Regression prevention measures</safeguards> </output_format>
Ready to paste
CodingGitHub Copilot

Multi-File Clean Refactor with Copilot Edits

Standard architectural pattern for decoupling database queries into reusable service functions using Copilot.

You are a Senior Systems Architect reviewing our TypeScript codebase. Context: The active file handles database operations directly inside API handlers. We need to separate concerns into our service layer. Task: Extract database logic into a dedicated function in @/lib/services, refactor the route handler to call it, and supply unit tests. Constraints: - Strictly adhere to existing project error handling conventions. - Maintain full type safety with zero 'any' types. - Ensure database instances are never exposed to client-side code.
Ready to paste
CodingCursor

Test-Driven Bug Fix with Minimal Diff

Guides Cursor to make surgical, minimal diffs with accompanying unit tests.

Review @lib/auth.ts and @app/api/auth/route.ts. Problem: We are encountering intermittent race conditions during concurrent session refresh requests. Task: Implement a mutual exclusion lock or idempotent token check to guarantee single-execution session updates. Constraints: - Make the smallest safe contiguous change (under 30 lines). - Do not introduce external dependencies. - Update @tests/auth.test.ts with a concurrent simulation test.
Ready to paste