fix(security): hard-disable litellm integrations

[qin-ctx]

Description

Temporarily hard-disable LiteLLM across the runtime, config, and dependency surface in response to the public supply-chain compromise report for recent LiteLLM releases.

Related Issue

External security incident: BerriAI/litellm#24512

Type of Change

• Bug fix (non-breaking change that fixes an issue)
• New feature (non-breaking change that adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation update
• Refactoring (no functional changes)
• Performance improvement
• Test update

Changes Made

• Remove LiteLLM from the root dependency and lockfile, and stop advertising it as a supported VLM or embedding provider.
• Fail fast when provider=litellm is configured for VLM, embedding, or rerank instead of falling through to runtime import errors.
• Disable LiteLLM-backed bot provider and image generation entry points, and replace old LiteLLM tests with hard-disable regression coverage.

Testing

• 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 tested this on the following platforms:
  • Linux
  • macOS
  • Windows

Checklist

• My code follows the project's coding style
• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• Any dependent changes have been merged and published

Screenshots (if applicable)

N/A

Additional Notes

Validated with:

• python -m pytest -o addopts='' tests/unit/test_litellm_embedder.py -q
• python -m pytest -o addopts='' tests/unit/test_extra_headers_vlm.py -k LiteLLMDisabled -q

I did not mark the full unit test suite as passing. An existing subset in tests/unit/test_extra_headers_vlm.py still has unrelated mock target failures outside the new LiteLLM-disable coverage.

[github-actions]

Failed to generate code suggestions for PR