Add Zscaler AI Guard hook

[jwang-gif]

Title

Pre-Submission checklist

Please complete all items before asking a LiteLLM maintainer to review your PR

• [ Y] I have Added testing in the tests/litellm/ directory, Adding at least 1 test is a hard requirement - see details
• [ Y] I have added a screenshot of my new test passing locally
• [ Y] My PR passes all unit tests on make test-unit
• [ Y] My PR's scope is as isolated as possible, it only solves 1 specific problem

Type

🆕 New Feature

Changes

-Add guard hook for Zscaler AI Guard (code, test, doc)
-Support moderation hooks(during-call, pre-call) and post-call

[vercel]

@jwang-gif is attempting to deploy a commit to the CLERKIEAI Team on Vercel.

A member of the Team first needs to authorize it.

[krrishdholakia]

Hi @jwang-gif can you fix your linting checks -

[jwang-gif]

hi @krrishdholakia & team
For the latest commit(9), the lint test pass ,
see unit test got canceled, attached the local run results for all guardrail related unit test.
the failed cases are from existing code, and they all failed on main branch, also look at the error message, they are not caused by the code change.
Screenshot for run on my branch.

Screenshot for run on main branch

[Vercel] failure not related to new change

[krrishdholakia]

please remove changes here. it should not be needed. the guardrail init should be handling this

[krrishdholakia]

same comment - please remove changes here, the guardrail init should be handling this

[krrishdholakia]

if you implement the apply_guardrail function, your guardrail will work across all endpoints with this PR: #15706