Build your next AI agent in Python, as a prompt or in n8n, starting from kits that already pass their tests.
OpenAI Agents SDK starter in Python: triage, orders and billing agents, handoffs, guardrails, human approval for refunds, sessions and 23 offline tests.
View OpenAI Agents SDK Starter: Customer Support Desk11 AI agent system prompt templates (support, triage, RAG, n8n, SQL, coding, research) with 16 test cases you can run on any model. All pass on a real model.
View AI Agent System Prompts Pack: 11 Tested Prompts4 n8n Telegram workflows: AI assistant bot with memory, voice-note transcriber, daily news digest, approval bot. Tested in n8n 2.46. $19.
View n8n Telegram AI Bot WorkflowsThree tested kits for building AI agents three ways: in Python code, as reusable system prompts, and as no-code n8n workflows. Each kit keeps its own README, tests and demo.
| Folder | What it is |
|---|---|
| kits/openai-agents-sdk-starter/ | Multi-agent customer-support app on the OpenAI Agents SDK: triage with handoffs to orders and billing agents, refunds over a limit paused for human approval, pause-and-resume, session memory, card and prompt-injection guardrails. 23 offline tests |
| kits/ai-agent-system-prompts/ | 11 system prompts (support, triage, RAG with citations, n8n AI Agent, SQL analyst, lead qualifier, meeting notes, review replies, coding agent, brand voice, research) with 16 test cases and promptcheck.py to run them on any OpenAI-compatible model. 14 tests |
| kits/n8n-telegram-ai-bot-workflows/ | 4 n8n workflows: Telegram AI assistant with memory and allow-list, voice-note transcriber, daily digest, approval bot. Lint-checked JSON plus tested Code nodes. 13 tests |
Bought separately: $39 + $19 + $19 = $77.
./run_tests.sh # every kit's tests, offline (the Agents SDK kit needs Python 3.10+ and pip install -r kits/openai-agents-sdk-starter/requirements.txt)
./demo.sh # offline demos of all threeThe screenshot is real output from the kit’s own commands. Tested from a fresh download of this exact zip: ./run_tests.sh passed on Python 3.13 and 3.10.
./run_tests.sh passes on Python 3.13 and 3.10 (openai-agents 0.24.0): 23 + 14 + 13 tests. The system prompts' 16 cases also pass on a real model (Qwen3.8-27B, local, temperature 0); outputs are in kits/ai-agent-system-prompts/recorded/.
The limits in each kit's README apply. The Agents SDK kit needs an OpenAI API key for real conversations (tests and demo run offline); the n8n workflows need your own Telegram bot token and model credentials in n8n.
Zip, 253 KB, v1.0, 74 files. You may use and modify this kit for your own work and for clients you serve, including paid client work. Not affiliated with or endorsed by OpenAI, Telegram or n8n.
AI agent templates: OpenAI Agents SDK support desk, 11 tested system prompts and 4 n8n Telegram AI bot workflows, 50 offline tests. $49 ($77 separately).
OpenAI Agents SDK starter in Python: triage, orders and billing agents, handoffs, guardrails, human approval for refunds, sessions and 23 offline tests.
View OpenAI Agents SDK Starter: Customer Support DeskFree system prompt templates for a support agent, ticket triage (JSON) and RAG with citations, plus a test runner for any OpenAI-compatible model.
View System Prompt Templates Starter (Free)Test your LLM app or agent against 119 prompt injection attacks: system prompt leaks, RAG injection, tool hijack, exfiltration, encodings. CI-ready.
View Prompt Injection Test Suite: 119 Test Cases for LLM Apps + CIShort weekly digest of new x402 + MCP tools and free kits, including the MCP OAuth Checker. No spam, unsubscribe any time.