Bug Description

When using custom_providers configuration in Hermes Agent, model switching via the --model parameter fails with "Model Not Exist" error from the API. The same configuration works perfectly when set as the default model.

Environment

+- Hermes Agent Version: v0.8.0 (2026.4.8) · upstream af9caec
+- OS: Linux
+- Python: 3.12

Steps to Reproduce

+### 1. Configuration
+Add to ~/.hermes/config.yaml:
+```yaml
+custom_providers:
+- name: qwen3.6-plus

• base_url: https://dashscope.aliyuncs.com/compatible-mode/v1
• api_key: ${ALI_API_KEY}
• api_mode: chat_completions
• model: qwen3.6-plus # Actual model name for the API
  +```

+### 2. Test with custom_providers (FAILS)
+bash +hermes chat --model qwen3.6-plus -q "Hello" --quiet +
+Result: Session created but no response. Log shows:
+ +ERROR root: Non-retryable client error: Error code: 400 - {'error': {'message': 'Model Not Exist', 'type': 'invalid_request_error', 'param': None, 'code': 'invalid_request_error'}} +
+
+### 3. Test as default model (WORKS)
+Change default config:
+```yaml
+model:

• default: qwen3.6-plus
• provider: custom
• base_url: https://dashscope.aliyuncs.com/compatible-mode/v1
• api_key: ${ALI_API_KEY}
  + +Then: +bash
  +hermes chat -q "Hello" --quiet
  +```
  +Result: Works perfectly, model responds correctly.

+### 4. Direct API test (WORKS)
+```bash
+curl -X POST https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions \

• -H "Authorization: Bearer $ALI_API_KEY" \
• -H "Content-Type: application/json" \
• -d '{"model": "qwen3.6-plus", "messages": [{"role": "user", "content": "Hello"}], "max_tokens": 50}'
  +```
  +Result: API responds correctly.

Expected Behavior

+When using --model qwen3.6-plus, Hermes should:
+1. Recognize qwen3.6-plus as a custom_provider
+2. Use the configured base_url, api_key, and model field
+3. Successfully call the API with model name qwen3.6-plus

Actual Behavior

1. Hermes recognizes qwen3.6-plus as a custom_provider
   +2. But sends an incorrect request causing "Model Not Exist" error
   +3. Same configuration works when set as default model

Affected Component

Configuration (config.yaml, .env, hermes setup)

Messaging Platform (if gateway-related)

No response

Operating System

Ubuntu24.04

Python Version

3.12

Hermes Version

0.8.0

Relevant Logs / Traceback

Root Cause Analysis (optional)

+After inspecting the code:
+
+### Problem 1: ProviderDef missing model field
+In hermes_cli/providers.py:
+```python
+class ProviderDef:

• id: str
• name: str
• transport: str
• api_key_env_vars: Tuple[str, ...]

Proposed Fix (optional)

No response

Are you willing to submit a PR for this?

• I'd like to fix this myself and submit a PR