Pipe Explorer (VS Code Extension)
Developed by Abhineet Pandey
Email: abhineet.pandey89@gmail.com
GitHub: abhineet-pandey
Pipe Explorer is a VS Code extension for managing long-running compute/service pipes from a single sidebar (Slurm + Jupyter, TensorFlow services, and custom commands).
Preview

Features
- Tree UI with clear sections:
Options
Active / Connected
All Pipes
- Green dot indicator for active/connected pipes.
- Command-based lifecycle:
- Secure secret handling through VS Code SecretStorage.
- SSH config host discovery from
~/.ssh/config (or custom path).
- JSON import/export with schema versioning.
- Slurm tunnel support with explicit node target (
slurmNodeHost).
- Two pipe setup modes:
Guided setup (field-by-field)
Full command (one-step) (paste complete command)
Commands
Pipe Explorer: Add Pipe
Pipe Explorer: Start Pipe
Pipe Explorer: Close Pipe
Pipe Explorer: Stop Pipe (compat alias)
Pipe Explorer: Set Pipe Secret
Pipe Explorer: Import Pipes
Pipe Explorer: Export Pipes
Pipe Explorer: Remove Pipe
Pipe Explorer: Refresh
Add Pipe setup modes
When running Add Pipe, choose one mode:
Guided setup
- asks for structured fields (type, SSH host, ports, Slurm settings, etc.)
Full command (one-step)
- asks only for pipe name + full command
- creates a
custom pipe with that command directly
Configuration
pipeManager.defaultSshHost
pipeManager.sshConfigPath
Slurm + Jupyter tunnel behavior
Default tunnel format:
ssh -N -L <localPort>:localhost:<remotePort> <sshHost>
If slurmNodeHost is set (recommended for multi-node clusters):
ssh -N -L <localPort>:<slurmNodeHost>:<remotePort> <sshHost>
Example:
ssh -N -L 8888:dgx01:8888 slurmgpuserver
Security
- Secret values are never written to export files.
secretKey is stored in pipe config; secret value is read from SecretStorage at runtime.
{
"version": 1,
"pipes": [
{
"id": "1700000000000-my-pipe",
"name": "my-pipe",
"type": "slurm-jupyter",
"sshHost": "user@cluster",
"workdir": "~",
"slurmPartition": "gpu",
"slurmGpus": 1,
"slurmTime": "02:00:00",
"slurmNodeHost": "dgx01",
"jupyterPort": 8888,
"jupyterLocalPort": 8888,
"autoTunnel": true,
"env": "~/miniconda3/bin/activate"
}
]
}
Build and run
npm install
npm run compile
Then press F5 in VS Code to launch Extension Development Host.
Documentation
- Architecture:
docs/ARCHITECTURE.md
- Operations:
docs/OPERATIONS.md