What does this PR do?

Fixes an issue where models selected via /model custom:yyyy are automatically overridden by the default_model from the provider configuration.
When users configure a custom provider (such as Ollama, vLLM, etc.) with a default_model, switching to a different model using the /model command would be overridden in the next conversation turn, resetting back to the default model.
Root Cause: The _ensure_runtime_credentials() method in cli.py calls resolve_runtime_provider() at the start of each conversation turn, which retrieves the provider configuration including the default_model. The original code unconditionally overwrote the user's model selection with this runtime model, without distinguishing between:

• User's explicitly chosen model (should be preserved)
• Provider name as model (needs resolution to default_model)
  Fix: Added conditional logic to only apply the runtime default_model when:

1. No model is set (first-time use)
2. The model equals the provider name (needs resolution to default_model)
3. The model matches the provider's display name
   This preserves the user's explicit model selection while maintaining the existing auto-resolution functionality.

Related Issue

Fixes # (fill in if there's a related issue)

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

• cli.py (lines 3149-3158): Modified model override logic in _ensure_runtime_credentials() method
  • Added conditional check should_use_runtime_model
  • Only applies runtime default_model when model is unset or equals provider name
  • Preserves user's explicit model selection via /model command

How to Test

Steps to reproduce the bug (before fix):

1. Configure a custom provider:

# ~/.hermes/config.yaml
providers:
  my-ollama:
    base_url: "http://localhost:11434/v1"
    default_model: "llama3.2"

2. Start Hermes CLI and switch model:

/model my-ollama:qwen2.5

3. Send a message
4. Model gets overridden to llama3.2 ❌
   Steps to verify the fix (after fix):
5. Repeat the above steps
6. After sending a message, model remains my-ollama:qwen2.5 ✅
   Run tests:

pytest tests/hermes_cli/test_user_providers_model_switch.py -v
pytest tests/run_agent/test_switch_model_* -v

All tests should pass (20/20).

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: Ubuntu 24.04

Documentation & Housekeeping

• I've updated relevant documentation (README, docs/, docstrings) — N/A (internal logic fix, no doc updates needed)
• I've updated cli-config.yaml.example if I added/changed config keys — N/A (no config changes)
• 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 — Change only involves Python logic, no platform-specific code
• I've updated tool descriptions/schemas if I changed tool behavior — N/A (no tool changes)

For New Skills

N/A (not adding a new skill)

Screenshots / Logs

Before fix:

User: /model my-ollama:qwen2.5
✓ Model switched: my-ollama:qwen2.5
User: (send message)
[internal log] Overriding model to llama3.2  ← Bug!

After fix:

User: /model my-ollama:qwen2.5
✓ Model switched: my-ollama:qwen2.5
User: (send message)
✓ Model remains my-ollama:qwen2.5  ← Fixed!

Test output:

tests/hermes_cli/test_user_providers_model_switch.py::test_switch_model_resolves_user_provider_credentials PASSED
============================== 15 passed in 1.21s ===============================