Fix adversarial chat target for Tense, Crescendo, and MultiTurn attack strategies

[slister1001]

Description

Please add an informative description that covers that changes made by the pull request and link all relevant issues.

If an SDK is being regenerated based on a new API spec, a link to the pull request containing these API spec changes should be included above.

All SDK Contribution checklist:

• The pull request does not introduce [breaking changes]
• CHANGELOG is updated for new features, bug fixes or other significant changes.
• I have read the contribution guidelines.

General Guidelines and Best Practices

• Title of the pull request is clear and informative.
• There are a small number of commits, each of which have an informative message. This means that previously merged commits do not appear in the history of the PR. For more information on cleaning up the commits in your PR, see this page.

Testing Guidelines

• Pull request includes test coverage for the included changes.

[Copilot]

Pull request overview

This PR fixes Foundry-based red team execution so PyRIT’s adversarial_chat_target uses an AzureRAIServiceTarget (instead of the user’s callback target), preventing callback responses from leaking into converted prompts and ultimately into scan result “user message” content.

Changes:

• Updated _execute_attacks_with_foundry to construct and pass an AzureRAIServiceTarget as adversarial_chat_target, selecting a template key based on attack strategies.
• Updated Foundry result message construction to show original_value for user messages (falling back to converted_value when needed).
• Added regression unit tests and a changelog entry covering the callback-response leakage scenario.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 4 comments.

File	Description
sdk/evaluation/azure-ai-evaluation/azure/ai/evaluation/red_team/_red_team.py	Ensures Foundry adversarial chat uses AzureRAIServiceTarget and adds strategy→template selection helper.
sdk/evaluation/azure-ai-evaluation/azure/ai/evaluation/red_team/_foundry/_foundry_result_processor.py	Displays the original adversarial prompt for user messages to avoid showing converter/callback output.
sdk/evaluation/azure-ai-evaluation/tests/unittests/test_redteam/test_foundry.py	Updates existing mocks for original_value and adds new regression tests around adversarial chat wiring and message building.
sdk/evaluation/azure-ai-evaluation/CHANGELOG.md	Documents the bug fix in the unreleased section.

---

You can also share your feedback on Copilot code review. Take the survey.

[nagkumar91]

Good bug fix. The root cause — user's callback being passed as adversarial_chat_target and leaking into converter output — is clearly identified and correctly fixed by creating a proper AzureRAIServiceTarget instead.

• _get_adversarial_template_key logic is sound — _validate_strategies already prevents Crescendo+MultiTurn coexistence, so the precedence concern is moot ✅
• original_value preference for user messages is a sensible defense-in-depth ✅
• crescendo_format flag properly gated on template selection ✅
• Tests are thorough: 9 regression tests covering template selection, wiring, original_value display, and the exact reported bug ✅

LGTM.