Python: Update hosting agent samples + fixes

[TaoChenOSU]

Motivation and Context

Current samples don't align with samples in other Azure repositories. There are also defects in the responses host.

Description

1. Update foundry hosting agent samples to align with the samples in other Azure repos.
2. Fix file content type support and add more tests.

Contribution Checklist

• The code builds clean without any errors or warnings
• The PR follows the Contribution Guidelines
• All unit tests pass, and I have added new tests where possible
• Is this a breaking change? If yes, add "[BREAKING]" prefix to the title of the PR.

[moonbox3]

Python Test Coverage Report •

File	Stmts	Miss	Cover	Missing
packages/foundry_hosting/agent_framework_foundry_hosting
_responses.py	524	114	78%	151–152, 161–162, 166, 171, 185, 209, 235–237, 256–258, 260–262, 264–266, 270–273, 276–281, 288, 293, 296, 303–304, 306–307, 309, 311, 313–316, 318–320, 324, 327–334, 337–338, 340–341, 349–354, 426–427, 1072–1074, 1076, 1123–1124, 1126–1127, 1129–1130, 1132–1133, 1138, 1147, 1150–1152, 1154, 1172, 1175, 1210–1214, 1218–1224, 1231–1235, 1241–1247, 1254, 1260, 1263
TOTAL	29883	3479	88%

Python Unit Test Overview

Tests	Skipped	Failures	Errors	Time
6029	30 💤	0 ❌	0 🔥	1m 42s ⏱️

[github-actions]

Automated Code Review

Reviewers: 4 | Confidence: 84%

✓ Correctness

The PR adds file_data support for the input_file content type in the Responses hosting layer, refactors samples (renaming env vars, switching to DefaultAzureCredential, reorganizing directories), adds integration test infrastructure for foundry_hosting, and updates CI workflows. The core logic in _convert_file_data and the input_file/input_image changes are correct: data URIs route through Content.from_uri which correctly classifies them as type data, text/* files are decoded to plain text, and the MessageContentInputFileContent model does have file_data and filename attributes (verified against source). No blocking correctness issues found.

✓ Security Reliability

The PR adds file_data (base64 data URI) support, integration tests, CI workflow changes, and sample updates. The main reliability concern is in _convert_file_data where base64.b64decode().decode('utf-8') can raise unhandled UnicodeDecodeError or binascii.Error on malformed input at the HTTP trust boundary. The remaining changes (env var renames, credential switch to DefaultAzureCredential, README updates, workflow configs) are straightforward and carry no security or reliability risk.

✓ Test Coverage

The new _convert_file_data function has two distinct code paths (text/* decode vs. URI passthrough), but the only unit test exercises the non-text branch (application/pdf). The text/* branch — which base64-decodes content and returns Content.from_text() — is covered only by integration tests that require live credentials. Additionally, the media_type="image/*" change to Content.from_uri for input_image has no unit-level verification. The integration test suite (test_responses_int.py) is thorough and well-structured.

✗ Design Approach

I found two design-level issues worth addressing before merging. The new file_data conversion path special-cases all text/* uploads by decoding them as UTF-8 text, which is a losy workaround on top of a content layer that already knows how to preserve data URIs as typed data content. Separately, the new MCP sample now silently drops its only remote tool when GITHUB_PAT is missing, which turns a misconfigured MCP sample into an apparently healthy plain-chat sample instead of failing fast.

Suggestions

• The media_type="image/*" argument to Content.from_uri for input_image prevents _validate_uri from extracting the specific media type (e.g. image/jpeg) from data URIs, because extraction only occurs when media_type is None (_types.py:247). Consider omitting media_type for data URIs or letting validation extract it automatically.

---

Automated review by TaoChenOSU's agents

[Copilot]

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.

[Copilot]

Pull request overview

Copilot reviewed 50 out of 63 changed files in this pull request and generated 7 comments.

Comments suppressed due to low confidence (2)

python/samples/04-hosting/foundry-hosted-agents/invocations/README.md:1

• This file is now empty. If the content was intentionally moved to the parent README, consider either deleting this README entirely or leaving a short pointer to the parent README so users who land here don’t get a blank page.
  python/samples/04-hosting/foundry-hosted-agents/responses/README.md:1
• This file is now empty. If the content was intentionally moved to the parent README, consider either deleting this README entirely or leaving a short pointer to the parent README so users who land here don’t get a blank page.

[github-actions]

Automated Code Review

Reviewers: 4 | Confidence: 89% | Result: All clear

Reviewed: Correctness, Security Reliability, Test Coverage, Design Approach

---

Automated review by TaoChenOSU's agents