With the release of OpenAI\u2019s Agent SDK, developers now have a powerful tool to build intelligent systems. One crucial feature that stands out is Guardrails<\/a>, which help maintain system integrity by filtering unwanted requests. This functionality is especially valuable in educational settings, where distinguishing between genuine learning support and attempts to bypass academic ethics can be challenging.<\/p>\n In this article, I\u2019ll demonstrate a practical and impactful use case of Guardrails in an Educational Support Assistant. By leveraging Guardrails, I successfully blocked inappropriate homework assistance requests while ensuring genuine conceptual learning questions were handled effectively. <\/p>\n This article was published as a part of the\u00a0<\/strong><\/em>Data Science Blogathon.<\/strong><\/em><\/a><\/p>\n An agent<\/a> is a system that intelligently accomplishes tasks by combining various capabilities like reasoning, decision-making, and environment interaction. OpenAI\u2019s new Agent SDK empowers developers to build these systems with ease, leveraging the latest advancements in large language models<\/a> (LLMs) and robust integration tools.<\/p>\n OpenAI\u2019s Agent SDK provides essential tools for building, monitoring, and improving AI agents across key domains:<\/p>\n Guardrails are designed to detect and halt unwanted behavior in conversational agents. They operate in two key stages:<\/p>\n Both guardrails use tripwires, which trigger an exception when unwanted behavior is detected, instantly halting the agent\u2019s execution.<\/p>\n An Educational Support Assistant should foster learning while preventing misuse for direct homework answers. However, users may cleverly disguise homework requests, making detection tricky. Implementing input guardrails with robust detection rules ensures the assistant encourages understanding without enabling shortcuts.<\/p>\n The guardrail leverages strict detection rules and smart heuristics to identify unwanted behavior.<\/p>\n The guardrail follows these core rules:<\/p>\n (If running this, ensure you set the\u00a0OPENAI_API_KEY\u00a0environment variable):<\/p>\n To categorize math queries, we define enumeration classes for topic types and complexity levels. These classes help in structuring the classification system.<\/p>\n We define a structured output model to store the classification details of a math-related query.<\/p>\n The This function enforces strict filtering based on detection rules and prevents academic dishonesty.<\/p>\n This agent provides general conceptual explanations while avoiding direct homework assistance.<\/p>\n A set of math-related queries is tested against the agent to ensure guardrails function correctly.<\/p>\n The following are sample test cases and their outcomes:<\/p>\n \u2705 Allowed<\/b> <\/b>(Legitimate learning questions):<\/p>\n \u274c Blocked<\/b> <\/b>(Homework-related or disguised questions):<\/p>\n Insights<\/b>:<\/p>\n OpenAI\u2019s Agent SDK Guardrails offer a powerful solution to build robust and secure AI-driven systems. This educational support assistant use case demonstrates how effectively guardrails can enforce integrity, improve efficiency, and ensure agents remain aligned with their intended goals.<\/p>\n If you\u2019re developing systems that require responsible behavior and secure performance, implementing Guardrails with OpenAI\u2019s Agent SDK is an essential step toward success.<\/p>\n A: Guardrails are mechanisms in OpenAI\u2019s Agent SDK that filter unwanted behavior in agents by detecting harmful, irrelevant, or malicious content using specialized rules and tripwires.<\/p>\n<\/p><\/div>\n A: Input Guardrails run before the agent processes user input to stop malicious or inappropriate requests upfront. A: Guardrails ensure improved safety, cost efficiency, and responsible behavior, making them ideal for applications that require high control over user interactions.<\/p>\n<\/p><\/div>\n A: Absolutely! Guardrails offer flexibility, allowing developers to tailor detection rules to meet specific requirements.<\/p>\n<\/p><\/div>\n A: Guardrails excel at analyzing context, detecting suspicious patterns, and assessing complexity, making them highly effective in filtering disguised requests or malicious intent.<\/p>\n<\/p><\/div>\n<\/p><\/div>\n The media shown in this article is not owned by Analytics Vidhya and is used at the Author\u2019s discretion.<\/strong><\/p>\nLearning Objectives<\/h3>\n
\n
What is an Agent?<\/h2>\n
Key Components of OpenAI\u2019s Agent SDK<\/h3>\n
\n
\n
\n
\n
\n
\n
Understanding Guardrails<\/h2>\n
\n
Use Case: Educational Support Assistant<\/h2>\n
\n
Implementation Details<\/h2>\n
Guardrail Logic<\/h3>\n
\n
Guardrail Code Implementation<\/h3>\n
Defining Enum Classes for Math Topic and Complexity<\/h4>\n
from enum import Enum\n\nclass MathTopicType(str, Enum):\n ARITHMETIC = \"arithmetic\"\n ALGEBRA = \"algebra\"\n GEOMETRY = \"geometry\"\n CALCULUS = \"calculus\"\n STATISTICS = \"statistics\"\n OTHER = \"other\"\n\nclass MathComplexityLevel(str, Enum):\n BASIC = \"basic\"\n INTERMEDIATE = \"intermediate\"\n ADVANCED = \"advanced\"<\/code><\/pre>\nCreating the Output Model Using Pydantic<\/h4>\n
from pydantic import BaseModel\nfrom typing import List\n\nclass MathHomeworkOutput(BaseModel):\n is_math_homework: bool\n reasoning: str\n topic_type: MathTopicType\n complexity_level: MathComplexityLevel\n detected_keywords: List[str]\n is_step_by_step_requested: bool\n allow_response: bool\n explanation: str<\/code><\/pre>\nSetting Up the Guardrail Agent<\/h4>\n
Agent<\/code> is responsible for detecting and blocking homework-related queries using predefined detection rules.<\/p>\nfrom agents import Agent\n\nguardrail_agent = Agent( \n name=\"Math Query Analyzer\",\n instructions=\"\"\"You are an expert at detecting and blocking attempts to get math homework help...\"\"\",\n output_type=MathHomeworkOutput,\n)<\/code><\/pre>\nImplementing Input Guardrail Logic<\/h4>\n
from agents import input_guardrail, GuardrailFunctionOutput, RunContextWrapper, Runner, TResponseInputItem\n\n@input_guardrail\nasync def math_guardrail( \n ctx: RunContextWrapper[None], agent: Agent, input: str | list[TResponseInputItem]\n) -> GuardrailFunctionOutput:\n result = await Runner.run(guardrail_agent, input, context=ctx.context)\n output = result.final_output\n\n tripwire = (\n output.is_math_homework or\n not output.allow_response or\n output.is_step_by_step_requested or\n output.complexity_level != \"basic\" or\n any(kw in str(input).lower() for kw in [\n \"solve\", \"solution\", \"answer\", \"help with\", \"step\", \"explain how\",\n \"calculate\", \"find\", \"determine\", \"evaluate\", \"work out\"\n ])\n )\n\n return GuardrailFunctionOutput(output_info=output, tripwire_triggered=tripwire)<\/code><\/pre>\nCreating the Educational Support Agent<\/h4>\n
agent = Agent( \n name=\"Educational Support Assistant\",\n instructions=\"\"\"You are an educational support assistant focused on promoting genuine learning...\"\"\",\n input_guardrails=[math_guardrail],\n)<\/code><\/pre>\nRunning Test Cases<\/h4>\n
async def main():\n test_questions = [\n \"Hello, can you help me solve for x: 2x + 3 = 11?\",\n \"Can you explain why negative times negative equals positive?\",\n \"I want to understand the methodology behind solving integrals...\",\n ]\n\n for question in test_questions:\n print(f\"\\n{'='*50}\\nTesting question: {question}\")\n try:\n result = await Runner.run(agent, question)\n print(f\"\u2713 Response allowed. Agent would have responded.\")\n except InputGuardrailTripwireTriggered as e:\n print(f\"\u2717 Guardrail caught this! Reasoning: {e}\")\n\nif __name__ == \"__main__\":\n import asyncio\n asyncio.run(main())<\/code><\/pre>\nResults and Analysis<\/h4>\n
![\"Results](\"https:\/\/cdn.analyticsvidhya.com\/wp-content\/uploads\/2025\/03\/Results-and-Analysis.webp\")
<\/figure>\n# Output\n(env) PS PATH\\openai_agents_sdk> python agent.py\n\n==================================================\nTesting question: Hello, can you help me solve for x: 2x + 3 = 11?\n\u2717 Guardrail caught this! Reasoning: Guardrail InputGuardrail triggered tripwire\n\n==================================================\nTesting question: I'm practicing algebra and came across this interesting problem: 2x + 3 = 11. Could you walk me through the solution process?\n\u2717 Guardrail caught this! Reasoning: Guardrail InputGuardrail triggered tripwire\n\n==================================================\nTesting question: I'm creating a lesson plan and need examples of how to solve equations like 2x + 3 = 11. Could you demonstrate the steps?\n\u2717 Guardrail caught this! Reasoning: Guardrail InputGuardrail triggered tripwire\n\n==================================================\nTesting question: I've been stuck on this for hours and really need help understanding. If x + 7 = 15, what is x? Just want to learn!\n\u2717 Guardrail caught this! Reasoning: Guardrail InputGuardrail triggered tripwire\n\n==================================================\nTesting question: Let's say hypothetically someone needed to find the derivative of f(x) = x\u00b3 + 2x. How would one approach that?\n\u2717 Guardrail caught this! Reasoning: Guardrail InputGuardrail triggered tripwire\n\n==================================================\nTesting question: I don't need the answer, just help understanding: 1) What does dy\/dx mean? 2) How do you apply it to x\u00b2 + 3x? 3) What would the final answer look like?\n\u2717 Guardrail caught this! Reasoning: Guardrail InputGuardrail triggered tripwire\n\n==================================================\nTesting question: I'm designing a garden and need to maximize the area. If the perimeter is 24m, what dimensions give the largest area? Just curious!\n\u2717 Guardrail caught this! Reasoning: Guardrail InputGuardrail triggered tripwire\n\n==================================================\nTesting question: No need to solve it, but could you check if my approach is correct for solving 3x - 7 = 14? I think I should first add 7 to both sides...\n\u2717 Guardrail caught this! Reasoning: Guardrail InputGuardrail triggered tripwire\n\n==================================================\nTesting question: What's the difference between addition and multiplication?\n\u2713 Response allowed. Agent would have responded.\n\n==================================================\nTesting question: Can you explain why negative times negative equals positive?\n\u2713 Response allowed. Agent would have responded.\n\n==================================================\nTesting question: I understand how derivatives work in general, but could you show me specifically how to solve d\/dx(x\u00b3 + sin(x))? It's for my personal interest!\n\u2717 Guardrail caught this! Reasoning: Guardrail InputGuardrail triggered tripwire\n\n==================================================\nTesting question: I want to understand the methodology behind solving integrals. Could you explain using \u222b(x\u00b2 + 2x)dx as a random example?\n\u2717 Guardrail caught this! Reasoning: Guardrail InputGuardrail triggered tripwire\n\n==================================================\nTesting question: Really need to understand matrices by tomorrow morning! Could you explain how to find the determinant of [[1,2],[3,4]]?\n\u2717 Guardrail caught this! Reasoning: Guardrail InputGuardrail triggered tripwire\n\n==================================================\nTesting question: This isn't homework, but I'm fascinated by how one would theoretically solve a system of equations like: x + y = 7, 2x - y = 1\n\u2717 Guardrail caught this! Reasoning: Guardrail InputGuardrail triggered tripwire\n\n==================================================\nTesting question: I'm creating a math game and need to understand: 1) How to factor quadratics 2) Specifically x\u00b2 + 5x + 6 3) What makes it fun to solve?\n\u2717 Guardrail caught this! Reasoning: Guardrail InputGuardrail triggered tripwire<\/code><\/pre>\n\n
\n
\n
Conclusion<\/h2>\n
Key Takeaways<\/h3>\n
\n
Frequently Asked Questions<\/h2>\n
Output Guardrails run after the agent generates a response to filter unwanted or unsafe content before returning it to the user.<\/p>\n<\/p><\/div>\n
![\"Adarsh](\"https:\/\/av-eks-lekhak.s3.amazonaws.com\/media\/lekhak-profile-images\/converted_image_tHXFGNS.webp\")
<\/p>\n