[ CLINIC ] // D4 · TOOL USE QUALITY
Is your AI agent using its tools wrong?
Tool use fails in three quiet ways: the wrong tool called confidently, invented arguments, and ignoring the result it just got back. Each is testable across a fixed set of scenarios.
Part of the 6-dimension, 18-test battery · risk profile: UNCOMMON
What goes wrong
Calls the wrong tool confidently
hallucinates parameters
ignores tool output and answers from prior.
// TEST YOUR AGENT
Three probes to run before you trust it
1tool-selection — the right tool for each scenario
2argument-validity — types, ranges, hallucinated params
3result-integration — does it use what the tool returned?
How the Clinic fixes it
Tool allowlists + schema-locked arguments
pre-call validation
mandatory result citation.
[ FAQ ] // ANSWERS
Why does my AI agent call the wrong tool or make up parameters?
No allowlist, no argument validation, no requirement to use the result. Schema-locked arguments and mandatory result citation fix it — a scan measures selection accuracy and argument validity.
Get tool graded on YOUR agent
The free scan runs the real battery against your agent and returns an honest A–F grade with the failure modes named. In 24 hours you go from guessing to knowing.