MICROSOFT AGENT FRAMEWORK · 0.16.0-rc.1
Demo110 — Tool Reliability
Compare two descriptions with the same read fixture.
Objectives
- Compare ambiguous and precise tool descriptions against observed invocation payloads.
Prepare the demo
Start at the root of the complete solution already built in Release. Course prerequisites: Demo060. This program uses live chat from the selected profile; no embeddings are required except for Demo020 --embed. Check connectivity and required capabilities before attributing failure to the scenario.
| Project | Role |
|---|---|
| demos/Demo110_ToolReliability/Demo110.App.csproj | App |
Launch from the root
dotnet run --project demos/Demo110_ToolReliability/Demo110.App.csproj -c Release --no-buildModes and procedure
The program automatically runs the vague process catalogue, then the precise find_invoice_by_exact_id catalogue. Both use INV-001 from a simulated fixture. There is no external billing service or refund command.
var tool = AIFunctionFactory.Create(FindInvoice, new AIFunctionFactoryOptions
{
Name = precise ? "find_invoice_by_exact_id" : "process",
Description = precise
? "Reads the simulated invoice fixture by exact ID. INV-001 exists. Unknown IDs return NotFound and no value. Read only; cannot issue refunds."
: "Processes a request."
Results and useful failures
For each pass, read the assistant declaration, then Host answer or No invocation observed. A correct fixture lookup returns 42 EUR. If the model skips the tool, no value is verified. Record arguments and calls in an instrumented campaign for a real comparison; this console displays the observed result.
These observations are expected from inspecting the indicated code revision. No new model call was made to write this page; exact generated wording is not a reference output.
Scope and limits
Historical options 1/2/3 and multiple dangerous tools are no longer exposed. Both variants may succeed or both may fail; neither establishes statistical superiority.
