What are the ways a single tool call can fail?

A tool call fails in one of three places: the model picks the wrong tool, it fills in arguments that do not match what the tool needs, or it misreads what comes back once the tool runs. Tool selection fails when two tools look similar enough from their descriptions that the model guesses, or when nothing fits well and it forces one anyway. Argument formation fails when the model invents a value it does not actually have, carries a value over from the wrong turn, or gets the schema slightly wrong, a string where the tool wants a number. Output handling fails when the result comes back malformed, empty, or stale, and the model treats it as a normal answer instead of a problem worth flagging.

These three fail differently in how visible they are. A wrong tool choice usually shows up fast, the response looks obviously off. A subtly wrong argument or a misread result can look fine and only surface several turns later.

keep reading

More on this.

Why did my agent call the wrong tool with invented arguments?The model matched your request to the tool whose description sounded closest, then filled a gap in the arguments with a plausible guess instead of a real value.answer →Can parallel tool calls break if one depends on the other running first?Yes. Running two tools at once removes the ordering their logic depended on, and the dependent one can execute before its precondition exists.answer →Does tool selection accuracy get worse as an agent gets more tools?Yes. Narrowing a five-tool list down to only what a query needs raised selection accuracy on medium-difficulty queries by 16 percentage points in one study.answer →Everything on tool callingThe definition, and every question answered under tool calling.concept →

Send us the traces you already emit.

book a demo