Below is a list of test prompts you can use for evaluating whether or not your agent can conduct basic tool-calling that would be useful for general tasks:

Agentic Function	Prompt
Write Tool	Generate a simple README.md that only has the text “Hello World” in it.
Verify ✅	Agent created README.md successfully.
Read Tool	Tell me what is inside @README.md
Verify ✅	Agent outputs shows Hello World.
Web Search Tool	Search the web and tell me the latest exchange rate between Singapore SGD and Malaysia Ringgit. Provide the link to the source material used.
Verify ✅	Agent outputs a numeral output like 3.158 along with a valid web link that reflects the same value.