Nodes Manager (Extensions) "No nodes found" when selecting nodes for install

### 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 expect to be able to install custom nodes using Extensions / Nodes Manager. Specifically:
1) launch ComfyUI
2) Either:

- click "Extensions" button at top of window, or
- open ComfyUI drop-down "Menu" -> Extensions

3) Select a node pack from those shown
4) Select "Install"
5) The pack installs and works.

### Actual Behavior

At step (4) of the Expected Behavior, the "Install" buttons gray out and animate, and a status bar pops up at the bottom showing a spinning activity graphics and the text "Installing 0 of 0". However, nothing is output to the logging panel and nothing happens. The installation animations just continue. ComfyUI is not stuck, though, as the "installation" can be cancelled and the Nodes Manager window closed with their appropriate "X" buttons.

**Other symptoms:**
(1) For some of the Nodes Packs, when the "Node Pack Info" sidebar opens, the "Nodes" section at the bottom says "**No nodes found** The pack's nodes either could not be parsed, or the pack is a frontend extension only and doesn't have any nodes." However, for any nodes pack, whether this message appears or the "Nodes" does show included nodes, the install fails the same way.

(2) When selecting from the Nodes Manager menu any of...
- All installed
- Updates available
- Conflicting
- Missing nodes

... the main area displays "**Error connecting to the Comfy Node Registry.** Please try again later."

### Steps to Reproduce

Install ComfyUI. Activate its virtual environment. Launch. Try to install a custom node via "Extensions" button or Menu->Extensions.

### Debug Logs

```powershell
run with python main.py --listen --port 8192 --disable-all-custom-nodes

Adding extra search path audio_encoders /home//AI/shared_models/audio_encoders
Adding extra search path checkpoints /home//AI/shared_models/checkpoints
Adding extra search path clip /home//AI/shared_models/clip
Adding extra search path clip_vision /home//AI/shared_models/clip_vision
Adding extra search path controlnet /home//AI/shared_models/controlnet
Adding extra search path detection /home//AI/shared_models/detection
Adding extra search path diffusion_models /home//AI/shared_models/diffusion_models
Adding extra search path embeddings /home//AI/shared_models/embeddings
Adding extra search path gligen /home//AI/shared_models/gligen
Adding extra search path hypernetworks /home//AI/shared_models/hypernetworks
Adding extra search path latent_upscale_models /home//AI/shared_models/latent_upscale_models
Adding extra search path loras /home//AI/shared_models/loras
Adding extra search path photomaker /home//AI/shared_models/photomaker
Adding extra search path qwen-tts /home//AI/shared_models/qwen-tts
Adding extra search path sam3 /home//AI/shared_models/sam3
Adding extra search path style_models /home//AI/shared_models/style_models
Adding extra search path text_encoders /home//AI/shared_models/text_encoders
Adding extra search path unet /home//AI/shared_models/unet
Adding extra search path upscale_models /home//AI/shared_models/upscale_models
Adding extra search path vae /home//AI/shared_models/vae
Adding extra search path vae_approx /home//AI/shared_models/vae_approx
WARNING: You need pytorch with cu130 or higher to use optimized CUDA operations.
Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_mxfp8', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_mxfp8', 'scaled_mm_nvfp4']}
Found comfy_kitchen backend triton: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
Found comfy_kitchen backend cuda: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_nvfp4']}
Checkpoint files will always be loaded safely.
Total VRAM 15916 MB, total RAM 32020 MB
pytorch version: 2.11.0+cu128
xformers version: 0.0.35
Set vram state to: NORMAL_VRAM
Device: cuda:0 NVIDIA GeForce RTX 4060 Ti : cudaMallocAsync
Using async weight offloading with 2 streams
Enabled pinned memory 28817.0
Using xformers attention
aimdo: src/control.c:68:INFO:comfy-aimdo inited for GPU: NVIDIA GeForce RTX 4060 Ti (VRAM: 15915 MB)
DynamicVRAM support detected and enabled
Python version: 3.12.13 (main, Apr 4 2026, 16:59:37) [GCC 13.3.0]
ComfyUI version: 0.18.1
comfy-aimdo version: 0.2.12
comfy-kitchen version: 0.2.8
ComfyUI frontend version: 1.42.8
[Prompt Server] web root: /home//AI/comfyui_pr/venv_pr/lib/python3.12/site-packages/comfyui_frontend_package/static
Asset seeder disabled
Skipping loading of custom nodes
Context impl SQLiteImpl.
Will assume non-transactional DDL.
Starting server

To see the GUI go to: http://0.0.0.0:8192
```

### Other

System is Linux Mint 22.3 x86_64 , kernel: 6.17.0-19-generic

Installing the old ComfyUI Manager custom node manually and using that to install and manage customs nodes works as expected.

edit: I did try searching for existing issues related to this behavior and these error messages before creating this one.