models not found when in subfolder

### Custom Node Testing

- [x] I have tried disabling custom nodes and the issue persists (see [how to disable custom nodes](https://docs.comfy.org/troubleshooting/custom-node-issues#step-1%3A-test-with-all-custom-nodes-disabled) if you need help)

### Expected Behavior

i use comfy-manager extension to download models. it downloads them to models/diffusion_models/WAN/WAN2.1... as example.
Those are not found by comfy. When i move the model into models/diffusion_models, then its found. This happened with all models/*

### Actual Behavior

the model is shown as not found and the workflow does not execute

### Steps to Reproduce

install comfy as in the getting started, then install comfy-manager. use any workflow sample, use the extension to download/install the missing models

### Debug Logs

```powershell
[START] Security scan
[ComfyUI-Manager] Using uv as Python module for pip operations.
[DONE] Security scan
## ComfyUI-Manager: installing dependencies done.
** ComfyUI startup time: 2025-11-11 12:16:34.703
** Platform: Linux
** Python version: 3.12.9 (main, Feb 5 2025, 01:31:18) [GCC 14.2.0]
** Python executable: /home/oli/comfy-env/bin/python3.12
** ComfyUI Path: /home/oli/comfy/ComfyUI
** ComfyUI Base Folder Path: /home/oli/comfy/ComfyUI
** User directory: /home/oli/comfy/ComfyUI/user
** ComfyUI-Manager config path: /home/oli/comfy/ComfyUI/user/default/ComfyUI-Manager/config.ini
** Log path: /home/oli/comfy/ComfyUI/user/comfyui.log
[START] Security scan
[DONE] Security scan
## ComfyUI-Manager: installing dependencies done.
** ComfyUI startup time: 2025-11-11 12:16:34.782
** Platform: Linux
** Python version: 3.12.9 (main, Feb 5 2025, 01:31:18) [GCC 14.2.0]
** Python executable: /home/oli/comfy-env/bin/python3.12
** ComfyUI Path: /home/oli/comfy/ComfyUI
** ComfyUI Base Folder Path: /home/oli/comfy/ComfyUI
** User directory: /home/oli/comfy/ComfyUI/user
** ComfyUI-Manager config path: /home/oli/comfy/ComfyUI/user/default/ComfyUI-Manager/config.ini
** Log path: /home/oli/comfy/ComfyUI/user/comfyui.log

Prestartup times for custom nodes:
0.1 seconds: /home/oli/comfy/ComfyUI/custom_nodes/comfyui-manager
0.3 seconds: /home/oli/comfy/ComfyUI/custom_nodes/ComfyUI-Manager

Checkpoint files will always be loaded safely.
Total VRAM 24560 MB, total RAM 31168 MB
pytorch version: 2.9.0+rocm6.4
Set: torch.backends.cudnn.enabled = False for better AMD performance.
AMD arch: gfx1100
ROCm version: (6, 4)
Set vram state to: NORMAL_VRAM
Device: cuda:0 Radeon RX 7900 XTX : native
Enabled pinned memory 29609.0
Using pytorch attention
Python version: 3.12.9 (main, Feb 5 2025, 01:31:18) [GCC 14.2.0]
ComfyUI version: 0.3.68
ComfyUI frontend version: 1.28.8
[Prompt Server] web root: /home/oli/comfy-env/lib/python3.12/site-packages/comfyui_frontend_package/static
### Loading: ComfyUI-Manager (V3.37.1)
[ComfyUI-Manager] network_mode: public
### ComfyUI Version: v0.3.68-15-g5ebcab3c | Released on '2025-11-10'
[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/model-list.json
[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/alter-list.json
[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/github-stats.json
[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/extension-node-map.json
[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/custom-node-list.json
[/home/oli/comfy/ComfyUI/custom_nodes/comfyui_controlnet_aux] | INFO -> Using ckpts path: /home/oli/comfy/ComfyUI/custom_nodes/comfyui_controlnet_aux/ckpts
[/home/oli/comfy/ComfyUI/custom_nodes/comfyui_controlnet_aux] | INFO -> Using symlinks: False
[/home/oli/comfy/ComfyUI/custom_nodes/comfyui_controlnet_aux] | INFO -> Using ort providers: ['CUDAExecutionProvider', 'DirectMLExecutionProvider', 'OpenVINOExecutionProvider', 'ROCMExecutionProvider', 'CPUExecutionProvider', 'CoreMLExecutionProvider']
/home/oli/comfy/ComfyUI/custom_nodes/comfyui_controlnet_aux/node_wrappers/dwpose.py:26: UserWarning: DWPose: Onnxruntime not found or doesn't come with acceleration providers, switch to OpenCV with CPU device. DWPose might run very slowly
warnings.warn("DWPose: Onnxruntime not found or doesn't come with acceleration providers, switch to OpenCV with CPU device. DWPose might run very slowly")

Import times for custom nodes:
0.0 seconds: /home/oli/comfy/ComfyUI/custom_nodes/websocket_image_save.py
0.0 seconds: /home/oli/comfy/ComfyUI/custom_nodes/comfyui-manager
0.0 seconds: /home/oli/comfy/ComfyUI/custom_nodes/comfyui-segment-anything-2
0.0 seconds: /home/oli/comfy/ComfyUI/custom_nodes/ComfyUI-Grounding
0.0 seconds: /home/oli/comfy/ComfyUI/custom_nodes/comfyui-kjnodes
0.1 seconds: /home/oli/comfy/ComfyUI/custom_nodes/ComfyUI-Manager
0.2 seconds: /home/oli/comfy/ComfyUI/custom_nodes/comfyui_controlnet_aux
0.2 seconds: /home/oli/comfy/ComfyUI/custom_nodes/comfyui-tensorops

Context impl SQLiteImpl.
Will assume non-transactional DDL.
No target revision found.
Starting server

To see the GUI go to: http://127.0.0.1:8188
FETCH ComfyRegistry Data: 5/105
FETCH ComfyRegistry Data: 10/105
[DEPRECATION WARNING] Detected import of deprecated legacy API: /scripts/ui.js. This is likely caused by a custom node extension using outdated APIs. Please update your extensions or contact the extension author for an updated version.
[DEPRECATION WARNING] Detected import of deprecated legacy API: /extensions/core/groupNode.js. This is likely caused by a custom node extension using outdated APIs. Please update your extensions or contact the extension author for an updated version.
[DEPRECATION WARNING] Detected import of deprecated legacy API: /scripts/ui/components/buttonGroup.js. This is likely caused by a custom node extension using outdated APIs. Please update your extensions or contact the extension author for an updated version.
[DEPRECATION WARNING] Detected import of deprecated legacy API: /scripts/ui/components/button.js. This is likely caused by a custom node extension using outdated APIs. Please update your extensions or contact the extension author for an updated version.
FETCH ComfyRegistry Data: 15/105
FETCH ComfyRegistry Data: 20/105
FETCH ComfyRegistry Data: 25/105
FETCH ComfyRegistry Data: 30/105
FETCH ComfyRegistry Data: 35/105
FETCH ComfyRegistry Data: 40/105
FETCH ComfyRegistry Data: 45/105
FETCH ComfyRegistry Data: 50/105
FETCH ComfyRegistry Data: 55/105
FETCH ComfyRegistry Data: 60/105
FETCH ComfyRegistry Data: 65/105
FETCH ComfyRegistry Data: 70/105
FETCH ComfyRegistry Data: 75/105
FETCH ComfyRegistry Data: 80/105
FETCH ComfyRegistry Data: 85/105
FETCH ComfyRegistry Data: 90/105
FETCH ComfyRegistry Data: 95/105
FETCH ComfyRegistry Data: 100/105
FETCH ComfyRegistry Data: 105/105
FETCH ComfyRegistry Data [DONE]
[ComfyUI-Manager] default cache updated: https://api.comfy.org/nodes
FETCH DATA from: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/custom-node-list.json [DONE]
[ComfyUI-Manager] All startup tasks have been completed.
got prompt
model weight dtype torch.bfloat16, manual cast: None
model_type FLUX
VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16
no CLIP/text encoder weights in checkpoint, the text encoder model will not be loaded.
Requested to load MochiTEModel_
loaded completely; 95367431640625005117571072.00 MB usable, 9083.39 MB loaded, full load: True
CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cuda:0, dtype: torch.float16
Requested to load LTXV
loaded completely; 13850.91 MB usable, 3667.90 MB loaded, full load: True
100%|███████████████████████████████████████████| 30/30 [00:23<00:00, 1.27it/s]
Requested to load VideoVAE
loaded completely; 6500.01 MB usable, 799.55 MB loaded, full load: True
Warning: Ran out of memory when regular VAE decoding, retrying with tiled VAE decoding.
Prompt executed in 53.04 seconds
got prompt
100%|███████████████████████████████████████████| 30/30 [00:23<00:00, 1.28it/s]
Warning: Ran out of memory when regular VAE decoding, retrying with tiled VAE decoding.
Prompt executed in 28.61 seconds
got prompt
100%|███████████████████████████████████████████| 30/30 [00:23<00:00, 1.28it/s]
Warning: Ran out of memory when regular VAE decoding, retrying with tiled VAE decoding.
Prompt executed in 28.59 seconds
got prompt
100%|███████████████████████████████████████████| 30/30 [00:23<00:00, 1.28it/s]
Warning: Ran out of memory when regular VAE decoding, retrying with tiled VAE decoding.
Prompt executed in 28.68 seconds
FETCH DATA from: /home/oli/comfy/ComfyUI/user/default/ComfyUI-Manager/cache/4245046894_model-list.json [DONE]
FETCH DATA from: /home/oli/comfy/ComfyUI/user/default/ComfyUI-Manager/cache/4245046894_model-list.json [DONE]
Install model 'Qwen 2.5 VL 7B Text Encoder (fp8_scaled)' from 'https://huggingface.co/Comfy-Org/Qwen-Image_ComfyUI/resolve/main/split_files/text_encoders/qwen_2.5_vl_7b_fp8_scaled.safetensors' into '/home/oli/comfy/ComfyUI/models/text_encoders/qwen/qwen_2.5_vl_7b_fp8_scaled.safetensors'
100%|███████████████████████████████████████| 9.38G/9.38G [01:25<00:00, 110MB/s]

[ComfyUI-Manager] Queued works are completed.
{'install-model': 1}

After restarting ComfyUI, please refresh the browser.
FETCH DATA from: /home/oli/comfy/ComfyUI/user/default/ComfyUI-Manager/cache/4245046894_model-list.json [DONE]
Install model 'Qwen-Image Diffusion Model (fp8_e4m3fn)' from 'https://huggingface.co/Comfy-Org/Qwen-Image_ComfyUI/resolve/main/split_files/diffusion_models/qwen_image_fp8_e4m3fn.safetensors' into '/home/oli/comfy/ComfyUI/models/diffusion_models/qwen-image/qwen_image_fp8_e4m3fn.safetensors'
100%|███████████████████████████████████████| 20.4G/20.4G [03:08<00:00, 109MB/s]

[ComfyUI-Manager] Queued works are completed.
{'install-model': 1}

After restarting ComfyUI, please refresh the browser.
FETCH DATA from: /home/oli/comfy/ComfyUI/user/default/ComfyUI-Manager/cache/4245046894_model-list.json [DONE]
Install model 'Qwen-Image-Lightning 4steps V1.0' from 'https://huggingface.co/lightx2v/Qwen-Image-Lightning/resolve/main/Qwen-Image-Lightning-4steps-V1.0.safetensors' into '/home/oli/comfy/ComfyUI/models/loras/qwen-image-lightning/Qwen-Image-Lightning-4steps-V1.0.safetensors'
100%|███████████████████████████████████████| 1.70G/1.70G [00:14<00:00, 115MB/s]

[ComfyUI-Manager] Queued works are completed.
{'install-model': 1}

After restarting ComfyUI, please refresh the browser.
FETCH DATA from: /home/oli/comfy/ComfyUI/user/default/ComfyUI-Manager/cache/4245046894_model-list.json [DONE]
Install model 'Qwen-Image-Lightning 4steps V1.0 (bf16)' from 'https://huggingface.co/lightx2v/Qwen-Image-Lightning/resolve/main/Qwen-Image-Lightning-4steps-V1.0-bf16.safetensors' into '/home/oli/comfy/ComfyUI/models/loras/qwen-image-lightning/Qwen-Image-Lightning-4steps-V1.0-bf16.safetensors'
100%|█████████████████████████████████████████| 850M/850M [00:07<00:00, 117MB/s]

[ComfyUI-Manager] Queued works are completed.
{'install-model': 1}

After restarting ComfyUI, please refresh the browser.
got prompt
Failed to validate prompt for output 60:
* CLIPLoader 75:38:
- Value not in list: clip_name: 'qwen_2.5_vl_7b_fp8_scaled.safetensors' not in ['qwen/qwen_2.5_vl_7b_fp8_scaled.safetensors', 't5/t5xxl_fp16.safetensors', 't5xxl_fp16.safetensors', 'umt5_xxl_fp8_e4m3fn_scaled.safetensors']
* LoraLoaderModelOnly 75:73:
- Value not in list: lora_name: 'Qwen-Image-Lightning-4steps-V1.0.safetensors' not in ['qwen-image-lightning/Qwen-Image-Lightning-4steps-V1.0-bf16.safetensors', 'qwen-image-lightning/Qwen-Image-Lightning-4steps-V1.0.safetensors']
* VAELoader 75:39:
- Value not in list: vae_name: 'qwen_image_vae.safetensors' not in ['wan_2.1_vae.safetensors', 'pixel_space']
Output will be ignored
invalid prompt: {'type': 'prompt_outputs_failed_validation', 'message': 'Prompt outputs failed validation', 'details': '', 'extra_info': {}}
```

### Other

_No response_

Reproduce the issue with ComfyUI-Manager by installing models into nested directories under models/diffusion_models, models/text_encoders, and models/loras. Start from the CLIPLoader, LoraLoaderModelOnly, and VAELoader validation messages in the log, then verify that nested model paths appear in the available lists and the workflow resolves them without moving files.

Written by the indexing model from the issue text.