[OsamaAnsar]

Link to Issue or Description of Change

1. Link to an existing issue (if applicable):

No existing issue found (searched open and closed issues and PRs for rest_api_tool, query parameter, array, explode).

2. Or, if no issue exists, describe the change:

Problem:
prepareRequestParams in core/src/tools/openapi_tool/rest_api_tool.ts does queryParams.append(originalName, String(argValue)) for every query parameter, so non-scalar values are mangled:

• an array [1, 2, 3] becomes ?id=1%2C2%2C3 (one comma-joined value)
• an object {a: 1} becomes ?filter=%5Bobject+Object%5D, which is [object Object]

OpenAPI 3.0 query parameters default to style: form, explode: true, which serializes an array as id=1&id=2&id=3. adk-python passes the list to httpx, which repeats the key the same way (checked with httpx 0.28.1: {'id': [1, 2, 3]} gives ?id=1&id=2&id=3). The existing parity tests only cover an array request body, not an array query parameter.

Solution:
A small appendQueryParam helper in rest_api_tool.ts:

• Array: one name=value pair per element. null and undefined elements are skipped and an empty array sends nothing (httpx does the same for an empty list).
• Plain object (also as an array element): JSON.stringify, instead of [object Object].
• Scalars: unchanged (String(value)).

Why JSON for objects: adk-python only gets a result here by accident (httpx sends the Python repr, filter=%7B%27a%27%3A+1%7D), so there is nothing to match. The spec default for an object would be a=1, which drops the parameter name and can collide with other parameters. JSON keeps the parameter name and is lossless. ApiParameter does not carry style/explode, so those are not honoured; this PR only makes the default (form, explode: true) correct and does not add other styles.

Path parameters are untouched (#638 is the separate path-encoding fix).

Testing Plan

Unit Tests:

• I have added or updated unit tests for my change.
• All unit tests pass locally.

Added a query parameter serialization block (8 tests) to core/test/tools/openapi_tool/rest_api_tool_test.ts.

Before/after, calling prepareRequestParams directly:

                 before                                         after
id=[1,2,3]       https://example.com/items?id=1%2C2%2C3         https://example.com/items?id=1&id=2&id=3
filter={a:1}     https://example.com/items?filter=%5Bobject+Object%5D   https://example.com/items?filter=%7B%22a%22%3A1%7D
flag=false       https://example.com/items?flag=false           https://example.com/items?flag=false

With only the rest_api_tool.ts change reverted, 7 of the 8 new tests fail (the 8th, the scalar regression guard, passes either way); with the change restored the whole core/test/tools/openapi_tool directory passes (19 files, 549 tests). prettier --check and eslint are clean on the touched files.

vitest run --project unit:core on my Windows machine: 5608 passed, 2 failed. The 2 failures (core/test/utils/file_utils_test.ts dangling symlink, core/test/cli/utils/logs_test.ts symlink) are EPERM: operation not permitted, symlink and fail the same way on an unmodified main; I did not run them on Linux or macOS. Because of this I left "All unit tests pass locally" unticked.

Manual End-to-End (E2E) Tests:

Not run against a live API. I only exercised prepareRequestParams (and thereby the URL it builds); I did not send a request to a real server, and did not run test:integration / test:e2e.

Checklist

• I have read the CONTRIBUTING.md document.
• I have performed a self-review of my own code.
• I have commented my code, particularly in hard-to-understand areas.
• I have added tests that prove my fix is effective or that my feature works.
• New and existing unit tests pass locally with my changes.
• I have manually tested my changes end-to-end.
• Any dependent changes have been merged and published in downstream modules.

Additional context

Behaviour change to be aware of: an array query value previously reached the server as one comma-joined value (id=1,2,3); it is now repeated keys. A server that relied on the comma form (style: form, explode: false) would need that style supported separately.

[Varun-S10]

Hi @OsamaAnsar, thank you for raising the PR to fix the issue. We have reviewed the code changes and it does resolve the issue. I will escalate the PR to the team to review, thank you for your patience and understanding.
Hi @kalenkevich, could you please take a look?