Evals silently fail in the adk-web frontend if the tool trajectory check fails

**Describe the bug**
Loading evaluation results for a particular eval case fails if the trajectory check does not pass.

**To Reproduce**
**Example Agent**
```
from google.adk.agents import Agent

def commerce_system(request: str) -> list[dict[str, str | int]]:
"""Responds to requests for 'promotions' data"""
if request.lower() == "promotions":
return [{"name": "Free delivery", "min_price": 50}]
else:
return []

prompt = """You are a shopping assistant that provides advice about **promotions**"""

root_agent = Agent(
name="root_agent",
model="gemini-2.5-flash",
instruction=prompt,
tools=[commerce_system],
)
```
**Example evalset**
```
{
"eval_set_id": "bug",
"eval_cases": [
{
"eval_id": "fails",
"conversation": [
{
"user_content": {
"parts": [
{
"text": "what promotions are there?"
}
],
"role": "user"
},
"final_response": {
"parts": [
{
"text": "There is a promotion for free delivery on orders over $50."
}
],
"role": "model"
},
"intermediate_data": {
"invocation_events": [
{
"author": "root_agent",
"content": {
"parts": [
{
"function_call": {
"args": {
"request": "PROMOTIONS"
},
"name": "commerce_system"
}
}
],
"role": "model"
}
},
{
"author": "root_agent",
"content": {
"parts": [
{
"function_response": {
"name": "commerce_system",
"response": {
"result": [
{
"name": "Free delivery",
"min_price": 50
}
]
}
}
}
],
"role": "user"
}
}
]
},
"creation_timestamp": 1760829983.020511
}
],
"session_input": {
"app_name": "app",
"user_id": "user"
},
"creation_timestamp": 1760830006.5009038
},
{
"eval_id": "works",
"conversation": [
{
"user_content": {
"parts": [
{
"text": "What promotions are there?"
}
],
"role": "user"
},
"final_response": {
"parts": [
{
"text": "There is a promotion for free delivery on orders over $50."
}
],
"role": "model"
},
"intermediate_data": {
"invocation_events": [
{
"author": "root_agent",
"content": {
"parts": [
{
"function_call": {
"args": {
"request": "promotions"
},
"name": "commerce_system"
}
}
],
"role": "model"
}
},
{
"author": "root_agent",
"content": {
"parts": [
{
"function_response": {
"name": "commerce_system",
"response": {
"result": [
{
"name": "Free delivery",
"min_price": 50
}
]
}
}
}
],
"role": "user"
}
}
]
},
"creation_timestamp": 1760829983.020511
}
],
"session_input": {
"app_name": "app",
"user_id": "user"
},
"creation_timestamp": 1760830006.5009038
}
],
"creation_timestamp": 1760829999.966551
}
```
**Expected behavior**
After executing the evaluations, you should see the session replayed with ❌ ✅ markers indicating success, with a comparison of the reference vs predicted.

**Actual behaviour**
The test case where the function args match precisely will be selectable, however test cases where the tools args do not match the expected value, will remain clickable but do nothing.

If the final response does not match, but the trajectory does pass, the eval case will load, however the tool calls will also be given ❌ ❌ indicators, despite being valid calls.

Stack trace in the browser:
```
main-FFZEKOTG.js:18 ERROR TypeError: e is not iterable
at t.formatToolUses (main-FFZEKOTG.js:3877:57149)
at t.addEvalFieldsToBotEvent (main-FFZEKOTG.js:3877:57770)
at t.addEvalCaseResultToEvents (main-FFZEKOTG.js:3877:57563)
at Object.next (main-FFZEKOTG.js:3877:58443)
at Yv.next (main-FFZEKOTG.js:14:3115)
at d0._next (main-FFZEKOTG.js:14:2796)
at d0.next (main-FFZEKOTG.js:14:2523)
at main-FFZEKOTG.js:14:17885
at zv._next (main-FFZEKOTG.js:14:5438)
at zv.next (main-FFZEKOTG.js:14:2523)
```

**Desktop (please complete the following information):**
- ADK version(pip show google-adk): 1.17 (and applies to past few versions)

**Model Information:**
- gemini-2.5-flash

**Additional Information**

Seems to be because the `intermediateData` does not populate `toolUses`, they are only represented in `intermediate_responses`