Skip to content
New this week26 prompts for Sora, Veo and Runway video workflows.
Explore video prompts→
AIPrompt123Prompt library

Search AIPrompt123

Search prompts, categories, platforms and repositories

26 prompts · reviewed by editors

AI Agents & Automation prompts

System design, tool schemas, evaluation and safety prompts for building reliable AI agents and automation workflows.

Agent system promptsTool schemasWorkflow designEvaluation rubrics

An agent is a loop: observe, decide, act, check. Most failures are not model failures; they are design failures. The agent has no stop condition, the tools are ambiguous, the state is lost between steps, or there is no verification before an irreversible action. The prompts in this category are engineering aids for the parts that are easy to skip: writing the system prompt, defining tool schemas, specifying hand-offs and building an evaluation set before you ship.

We write for people building with the OpenAI, Anthropic and Google APIs, workflow tools such as n8n and Zapier, and local orchestration frameworks. A good agent prompt names the role, the available tools, the decision rules, the stop conditions and the escalation path. It also tells the model what not to do. For tool schemas, clarity beats cleverness: one job per tool, explicit required fields, and errors that tell the model how to recover.

Prompts cover agent system prompts, planner-executor designs, tool definitions, retrieval workflows, human-in-the-loop checkpoints, evaluation rubrics, red-team tests and incident reviews. Start with a narrow workflow and a measurable success criterion. Do not give an agent write access to production systems without a confirmation step. The safest automation is the one you can pause, inspect and roll back.

All 26 ai agents & automation prompts

Filter in library →
Advanced

Write an agent system prompt with stop conditions

This copy-ready agents prompt helps you write an agent system prompt with stop conditions with fewer blank-page decisions. Paste it into ChatGPT and Claude, fill the variables, and you get a structured first draft you can edit. It is written for repeatable results and tells you what to check before publishing.

ChatGPTClaudeGemini
4.7(59)186 copies
Advanced

Define a tool schema the model can use correctly

A agents prompt for define a tool schema the model can use correctly. It gives ChatGPT and Claude a clear role, explicit constraints and a defined output, so the first answer is closer to usable. Fill in the variables, run it once, then use the verification notes on this page to review the result.

ChatGPTClaude
4.8(60)666 copies
Advanced

Design a planner-executor agent workflow

Design a planner-executor agent workflow with a prompt built for Claude and ChatGPT. It supplies context, constraints and an output contract, then asks the model to surface assumptions and gaps. Copy it, fill the variables and treat the result as a structured draft, not a final answer.

ClaudeChatGPT
4.7(14)532 copies
Advanced

Build an evaluation set for an agent

This copy-ready agents prompt helps you build an evaluation set for an agent with fewer blank-page decisions. Paste it into Claude and ChatGPT, fill the variables, and you get a structured first draft you can edit. It is written for repeatable results and tells you what to check before publishing.

ClaudeChatGPT
4.8(45)143 copies
Advanced

Red-team an agent for prompt injection

A agents prompt for red-team an agent for prompt injection. It gives Claude and ChatGPT a clear role, explicit constraints and a defined output, so the first answer is closer to usable. Fill in the variables, run it once, then use the verification notes on this page to review the result.

ClaudeChatGPT
4.8(60)1.0k copies
Intermediate

Design a human-in-the-loop approval step

Design a human-in-the-loop approval step with a prompt built for Claude and ChatGPT. It supplies context, constraints and an output contract, then asks the model to surface assumptions and gaps. Copy it, fill the variables and treat the result as a structured draft, not a final answer.

ClaudeChatGPT
4.8(33)222 copies
Advanced

Write a retrieval prompt grounded in sources

This copy-ready agents prompt helps you write a retrieval prompt grounded in sources with fewer blank-page decisions. Paste it into Claude and ChatGPT, fill the variables, and you get a structured first draft you can edit. It is written for repeatable results and tells you what to check before publishing.

ClaudeChatGPT
4.7(62)616 copies
Intermediate

Design an n8n automation flow

A agents prompt for design an n8n automation flow. It gives ChatGPT and Claude a clear role, explicit constraints and a defined output, so the first answer is closer to usable. Fill in the variables, run it once, then use the verification notes on this page to review the result.

ChatGPTClaude
4.8(63)1.1k copies
Advanced

Choose between a workflow and an agent

Choose between a workflow and an agent with a prompt built for Claude and ChatGPT. It supplies context, constraints and an output contract, then asks the model to surface assumptions and gaps. Copy it, fill the variables and treat the result as a structured draft, not a final answer.

ClaudeChatGPT
4.9(79)721 copies
Advanced

Write a multi-agent hand-off protocol

This copy-ready agents prompt helps you write a multi-agent hand-off protocol with fewer blank-page decisions. Paste it into Claude and ChatGPT, fill the variables, and you get a structured first draft you can edit. It is written for repeatable results and tells you what to check before publishing.

ClaudeChatGPT
4.8(75)705 copies
Advanced

Design an agent memory strategy

A agents prompt for design an agent memory strategy. It gives Claude and ChatGPT a clear role, explicit constraints and a defined output, so the first answer is closer to usable. Fill in the variables, run it once, then use the verification notes on this page to review the result.

ClaudeChatGPT
4.7(29)1.6k copies
Advanced

Write a guardrail checklist for production

Write a guardrail checklist for production with a prompt built for Claude and ChatGPT. It supplies context, constraints and an output contract, then asks the model to surface assumptions and gaps. Copy it, fill the details and treat the result as a structured draft, not a final answer.

ClaudeChatGPT
4.7(74)248 copies
Advanced

Design a structured extraction pipeline

This copy-ready agents prompt helps you design a structured extraction pipeline with fewer blank-page decisions. Paste it into ChatGPT and Claude, fill the variables, and you get a structured first draft you can edit. It is written for repeatable results and tells you what to check before publishing.

ChatGPTClaudeDeepSeek
4.7(56)158 copies
Intermediate

Write a customer support agent prompt

A agents prompt for write a customer support agent prompt. It gives ChatGPT and Claude a clear role, explicit constraints and a defined output, so the first answer is closer to usable. Fill in the variables, run it once, then use the verification notes on this page to review the result.

ChatGPTClaude
4.7(29)463 copies
Intermediate

Build a weekly research agent brief

Build a weekly research agent brief with a prompt built for Claude and Perplexity. It supplies context, constraints and an output contract, then asks the model to surface assumptions and gaps. Copy it, fill the variables and treat the result as a structured draft, not a final answer.

ClaudePerplexity
4.9(85)779 copies
Advanced

Design an agent cost model

This copy-ready agents prompt helps you design an agent cost model with fewer blank-page decisions. Paste it into Claude and ChatGPT, fill the variables, and you get a structured first draft you can edit. It is written for repeatable results and tells you what to check before publishing.

ClaudeChatGPTDeepSeek
4.8(15)1.7k copies
Intermediate

Write an escalation policy for an AI assistant

A agents prompt for write an escalation policy for an AI assistant. It gives Claude and ChatGPT a clear role, explicit constraints and a defined output, so the first answer is closer to usable. Fill in the variables, run it once, then use the verification notes on this page to review the result.

ClaudeChatGPT
4.8(36)274 copies
Advanced

Test an agent for instruction drift

Test an agent for instruction drift with a prompt built for Claude and ChatGPT. It supplies context, constraints and an output contract, then asks the model to surface assumptions and gaps. Copy it, fill the details and treat the result as a structured draft, not a final answer.

ClaudeChatGPT
4.7(23)925 copies
Advanced

Design a document review agent

This copy-ready agents prompt helps you design a document review agent with fewer blank-page decisions. Paste it into Claude and ChatGPT, fill the variables, and you get a structured first draft you can edit. It is written for repeatable results and tells you what to check before publishing.

ClaudeChatGPT
4.7(44)834 copies
Advanced

Write an agent observability plan

A agents prompt for write an agent observability plan. It gives Claude and ChatGPT a clear role, explicit constraints and a defined output, so the first answer is closer to usable. Fill in the context, run it once, then use the verification notes on this page to review the result.

ClaudeChatGPT
4.7(80)402 copies
Advanced

Design a content moderation agent

Design a content moderation agent with a prompt built for Claude and ChatGPT. It supplies context, constraints and an output contract, then asks the model to surface assumptions and gaps. Copy it, fill the variables and treat the result as a structured draft, not a final answer.

ClaudeChatGPT
4.9(16)1.4k copies
Intermediate

Write an agent onboarding checklist

This copy-ready agents prompt helps you write an agent onboarding checklist with fewer blank-page decisions. Paste it into Claude and ChatGPT, fill the details, and you get a structured first draft you can edit. It is written for repeatable results and tells you what to check before publishing.

ClaudeChatGPT
4.8(45)799 copies
Advanced

Design a browser automation agent

A agents prompt for design a browser automation agent. It gives Claude and ChatGPT a clear role, explicit constraints and a defined output, so the first answer is closer to usable. Fill in the variables, run it once, then use the verification notes on this page to review the result.

ClaudeChatGPT
4.7(44)978 copies
Advanced

Write a routing prompt for a multi-model system

Write a routing prompt for a multi-model system with a prompt built for Claude and ChatGPT. It supplies context, constraints and an output contract, then asks the model to surface assumptions and gaps. Copy it, fill the variables and treat the result as a structured draft, not a final answer.

ClaudeChatGPT
4.8(75)633 copies
Advanced

Review an agent's failure log

This copy-ready agents prompt helps you review an agent's failure log with fewer blank-page decisions. Paste it into Claude and ChatGPT, fill the variables, and you get a structured first draft you can edit. It is written for repeatable results and tells you what to check before publishing.

ClaudeChatGPT
4.8(81)827 copies
Intermediate

Design an AI workflow for a small team

A agents prompt for design an AI workflow for a small team. It gives Claude and ChatGPT a clear role, explicit constraints and a defined output, so the first answer is closer to usable. Fill in the variables, run it once, then use the verification notes on this page to review the result.

ClaudeChatGPT
4.9(79)1.3k copies
Verified Knowledge · FAQ

AI Agents & Automation prompt FAQ

What this collection covers, which models it supports and how to get a usable draft on the first try.