fix(agent): honor explicit api_mode on fallback chain entries

[kuziTony]

Fallback activation re-derived the wire protocol from provider name / base URL / model and ignored an api_mode the user had declared on the chain entry, so a fallback could silently come up on a different protocol than the one configured.

The primary path already honors an explicit value via runtime_provider._parse_api_mode; the fallback path had no equivalent. _iter_fallback_entries already passes the whole entry through, so the field was present in fb all along -- just never read.

Concretely, a self-hosted OpenAI-compatible gateway fronting Claude:

fallback_providers:
  - provider: custom model: claude-opus-4 base_url: http://gateway.internal:3000/v1 api_mode: anthropic_messages

matches none of the anthropic detection branches (provider is not anthropic, the URL neither ends in /anthropic nor resolves to api.anthropic.com), so inference fell through to chat_completions. Requests then went out over /chat/completions, which drops the cache_control blocks the Anthropic path attaches -- zeroing prompt caching for the rest of the conversation with nothing logged.

Observed on a self-hosted gateway: once traffic shifted to the fallback, session_model_usage recorded 77 calls with cache_read_tokens = 0, while the primary model on the same endpoint logged 274 calls / 18.4M cache_read_tokens.

Reuses _parse_api_mode for validation, so invalid or absent values fall through to the existing inference chain unchanged. Hostname detection added in #32243 / #49247 keeps working.

What does this PR do?

Related Issue

Fixes #

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

How to Test

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
• I've added tests for my changes (required for bug fixes, strongly encouraged for features)
• I've tested on my platform:

Documentation & Housekeeping

• I've updated relevant documentation (README, docs/, docstrings) — or N/A
• I've updated cli-config.yaml.example if I added/changed config keys — or N/A
• I've updated CONTRIBUTING.md or AGENTS.md if I changed architecture or workflows — or N/A
• I've considered cross-platform impact (Windows, macOS) per the compatibility guide — or N/A
• I've updated tool descriptions/schemas if I changed tool behavior — or N/A

For New Skills

• This skill is broadly useful to most users (if bundled) — see Contributing Guide
• SKILL.md follows the standard format (frontmatter, trigger conditions, steps, pitfalls)
• No external dependencies that aren't already available (prefer stdlib, curl, existing Hermes tools)
• I've tested the skill end-to-end: hermes --toolsets skills -q "Use the X skill to do Y"

Screenshots / Logs