fix(providers): complete native Gemini ListModels discovery

[LordMelkor]

What does this PR do?

Gemini live discovery now traverses the complete native ListModels catalog and keeps only models the chat picker can use. Native requests authenticate with x-goog-api-key, follow continuation tokens, require generateContent, and remove embedding, TTS, live, image, music, video, transcription, and other non-conversational families. A repeated continuation token fails discovery instead of returning a partial catalog.

Non-native endpoints retain the existing ProviderProfile.fetch_models() Bearer and data[].id contract, including the Google OpenAI-compatible endpoint and custom relays.

This carries forward the two commits from @PRATHAMESH75 in #62267 and adds the missing capability filtering and complete pagination behavior. It is intended to supersede #62267 rather than compete with it.

Related Issue

Fixes #62259

Supersedes #62267.

Type of Change

• 🐛 Bug fix (non-breaking change that fixes an issue)
• ✨ New feature (non-breaking change that adds functionality)
• 🔒 Security fix
• 📝 Documentation update
• ✅ Tests (adding or improving test coverage)
• ♻️ Refactor (no behavior change)
• 🎯 New skill (bundled or hub)

Changes Made

• plugins/model-providers/gemini/__init__.py — add native ListModels discovery with header authentication, complete pagination, capability filtering, and non-native endpoint delegation.
• tests/providers/test_fetch_models_base_url.py — exercise native discovery through a local HTTP server, including two-page traversal, filtering, repeated-token failure, missing credentials, and custom relay delegation.

How to Test

1. Run scripts/run_tests.sh tests/providers/test_fetch_models_base_url.py tests/plugins/model_providers/test_gemini_profile.py.
2. Confirm all 11 tests pass.
3. The new contract tests fail twice on the predecessor implementation: it neither performs the filtered native request nor delegates every non-native endpoint.

Checklist

Code

• I've read the Contributing Guide
• My commit messages follow Conventional Commits (fix(scope):, feat(scope):, etc.)
• I searched for existing PRs to make sure this isn't a duplicate
• My PR contains only changes related to this fix/feature (no unrelated commits)
• I've run pytest tests/ -q and all tests pass — focused suites passed through the repository's scripts/run_tests.sh wrapper
• I've added tests for my changes (required for bug fixes, strongly encouraged for features)
• I've tested on my platform: macOS 26.6.2, Python 3.11.15

Documentation & Housekeeping

• I've updated relevant documentation (README, docs/, docstrings) — N/A; the provider docstring describes the behavior
• I've updated cli-config.yaml.example if I added/changed config keys — N/A
• I've updated CONTRIBUTING.md or AGENTS.md if I changed architecture or workflows — N/A
• I've considered cross-platform impact (Windows, macOS) per the compatibility guide; the changed paths use standard-library HTTP and existing cross-platform helpers
• I've updated tool descriptions/schemas if I changed tool behavior — N/A

Screenshots / Logs

Focused validation: 11 passed, 0 failed. The Windows footgun scanner reported no findings in the two changed files.