Add comprehensive Istio list evaluation test

[josunect]

Summary

Adds a new MCP evaluation test istio-list-comprehensive that creates a larger set of Istio resources (31 total) to validate token efficiency at scale.

Motivation

Current evaluation tests (istio-list, istio-list-destination-rules) create only 1-3 Istio resources, which is insufficient to demonstrate the token savings of grouped output formats when listing many resources.

This test establishes a baseline that will show the true benefit of the namespace→GVK→{valid,invalid} grouped format vs the flat array format.

What's Created

The test creates in the bookinfo namespace:

• 10 VirtualServices (service-1-vs through service-10-vs)
• 10 DestinationRules (service-1-dr through service-10-dr) with v1/v2 subsets
• 5 Gateways (gateway-1 through gateway-5)
• 5 ServiceEntries (external-service-1 through external-service-5)
• 1 Invalid VirtualService (invalid-vs) to test validation error detection

Total: 31 Istio configuration objects

Expected Impact

With 31 resources, the grouped format should show:

• ~60-75% token reduction in the istio list output compared to flat array
• Demonstrates scalability benefits of deduplicating namespace/group/version/kind metadata

Test Details

• File: tests/evals/tasks/istio-list-comprehensive/istio-list-comprehensive.yaml
• Setup: Creates resources via kubectl apply
• Cleanup: Deletes all resources with label gevals.kiali.io/test=gevals-testing
• Verification: Checks that VirtualService, DestinationRule, and Gateway are mentioned in response

Next Steps

After merge:

1. CI will run mcpchecker and establish new baseline
2. PR [AI] Proposal for token reduction kiali/kiali#9945 (Istio list optimization) will show significant token savings on this test
3. Validates that grouped format scales better with realistic resource counts

---

🤖 Generated with Claude Code