AG-013Agent · Taking the pathlayaeval.NavigationEfficiency

Navigation Efficiency

Did the agent take the short way, without loops?

Simulation state:

When it passes

Direct path.

Threshold satisfied: evaluation score >= 3.0.

When it fails

Wandered in circles.

Violation detected: execution flagged for human review or fallback routing.

SDK Specification

Parameter	Value	Description
Class	NavigationEfficiency	Python evaluator class in layaeval
Required Fields	"task", "agent_trace", "expected_workflow"	Arguments expected in the call signature
Default Threshold	>= 3.0	Passes when quality score is at or above this value
Category	Agent Evaluation	Taking the path (Sensible order, no repeats, no going in circles.)

Python Example

pip install layaeval

navigation-efficiency.py

1from layaeval import NavigationEfficiency
2
3metric = NavigationEfficiency(threshold=3.0)  # passes at or above 3.0
4
5res = metric(
6    task="Book a hotel in Paris for 2 guests on Mar 4.",
7    agent_trace="Step 1: search_hotels(city='Paris') -> 3 results\nStep 2: book(id=12) -> confirmed",
8    expected_workflow="search -> choose -> book",
9)
10
11print(res.passed, res.value, res.confidence)

Related Agent Checks

Other checks in Taking the path

View all 42 checks →

AG-011Read →

Tool Redundancy

Did the agent ask for the same thing twice?

AG-012Read →

Tool Sequence

Were the tool steps done in a sensible order?