← All skills
remote-gpu-trainer
59 / 100 · Review before installing
Check before installing. Some of what it can do is expected: a coding skill usually runs build and test commands.
By sickn33/agentic-awesome-skills (GitHub) · coding · 49,524 stars · scanned at commit 00f9c40
Deploy, monitor, and debug long GPU jobs on RENTED/remote instances (AutoDL, RunPod, vast.ai, Lambda, Slurm, K8s): teardown/billing safety, spot resilience, resumable checkpointing, OOM/NaN triage.
Works with any agent that supports SKILL.md skills.
What this skill can do
Downloads and installs other software
The author does not mention this.
Found in 35 places.
Show 17 more
- skills/remote-gpu-trainer/profiles/autodl.md, line 211
and `apt-get install tmux` fails when egress is down. Zero-dependency fallback: - skills/remote-gpu-trainer/profiles/autodl.md, line 290
new-arch-capable torch (sm_120); the project pins `torch<2.9`; a naive `pip install -r requirements.txt` - skills/remote-gpu-trainer/profiles/china.md, line 264
Symptom: after `export …turbo.gpushare.com…`, `pip install` / `apt` / domestic mirrors hang or `ProxyError`. - skills/remote-gpu-trainer/profiles/generic-ssh.md, line 224
- **No root, shared multi-tenant node:** cannot `apt install`. Use `module load cuda` or a container - skills/remote-gpu-trainer/profiles/generic-ssh.md, line 418
run is glacial (ran on CPU) or fails to `pip install`/download. → Root cause: the **accelerator and - skills/remote-gpu-trainer/profiles/lambda.md, line 56
- **SkyPilot** — de-facto orchestration layer: `pip install "skypilot[lambda]"`, key file at - skills/remote-gpu-trainer/profiles/lambda.md, line 178
AutoDL tmux pattern. Install if absent (`sudo apt install -y tmux`); fall back to - skills/remote-gpu-trainer/profiles/lambda.md, line 231
- **LAM7 — `pip install torch` over Lambda Stack silently shadows or mismatches it.** - skills/remote-gpu-trainer/profiles/lambda.md, line 232
Symptom: a `pip install` in `base` reports *"Defaulting to user installation because normal site-packages - skills/remote-gpu-trainer/profiles/lambda.md, line 241
Symptom: created a conda env to use the Stack's torch, then a later `pip install` pulls a second, - skills/remote-gpu-trainer/profiles/lambda.md, line 275
- **LAM13 — GH200 (ARM/aarch64) breaks `pip install torch` — needs the ARM build.** - skills/remote-gpu-trainer/profiles/lambda.md, line 276
Symptom: on a 1× GH200 box, `pip install torch` installs a **CPU-only** wheel (no CUDA), or a pinned - skills/remote-gpu-trainer/profiles/lambda.md, line 279
from the CUDA index `pip install torch --index-url https://download.pytorch.org/whl/cu128` (aarch64 GPU - skills/remote-gpu-trainer/profiles/lambda.md, line 316
DETACH= tmux (apt install if absent; nohup fallback) - skills/remote-gpu-trainer/profiles/lambda.md, line 325
- Do not `pip install torch` / blanket `apt full-upgrade` on the rental — use the Stack as-is (LAM7/8/11); - skills/remote-gpu-trainer/profiles/paperspace.md, line 60
reproduces it. **On Notebooks, a plain `pip install` does NOT survive a restart** (writes to - skills/remote-gpu-trainer/profiles/paperspace.md, line 238
- **PS3 — `pip install` (or any non-`/storage` write) vanishes after a Notebook restart.**
and 15 more places
Sends data to these websites: hf-mirror.com
This is disclosed by the author in the skill's SKILL.md or README.
Found in 4 places.
Show 1 more
Reads your saved credentials, such as SSH keys or .env files
This is disclosed by the author in the skill's SKILL.md or README.
Found in 9 places.
Show 6 more
Changes files outside its own folder
The author does not mention this.
Found in 3 places.
Runs commands on your computer
This is disclosed by the author in the skill's SKILL.md or README. Expected for a coding skill.
Found in 1 place.
Safer alternatives
Low-risk skills in the same category (coding) that look related. Check that one does what you need.
Other copies in this repository
sickn33/agentic-awesome-skills has 2 more folders with a skill named remote-gpu-trainer. This page shows the lowest-scoring copy.
Install
Paste this into your terminal. It downloads exactly the version we scanned, not whatever is newest.
if [ -e ~/.agents/skills/'remote-gpu-trainer' ] || [ -L ~/.agents/skills/'remote-gpu-trainer' ]; then echo 'A skill folder named remote-gpu-trainer already exists — not replacing it' >&2; false; else d=$(mktemp -d) && git clone --quiet --filter=blob:none --no-checkout 'https://github.com/sickn33/agentic-awesome-skills' "$d" && git -C "$d" checkout --quiet '00f9c4077878be6ffcc093c099f05a6cccaf1b95' -- 'skills/remote-gpu-trainer' && mkdir -p ~/.agents/skills && cp -R "$d"/'skills/remote-gpu-trainer' ~/.agents/skills/'remote-gpu-trainer' && rm -rf "$d"; fi
Codex reads ~/.agents/skills directly. For Claude Code, Cursor, Copilot, Gemini CLI, Antigravity or OpenCode, also link it. For Claude Code: ln -s ~/.agents/skills/'remote-gpu-trainer' ~/.claude/skills/'remote-gpu-trainer' (other agents: ~/.cursor/skills, ~/.copilot/skills, ~/.gemini/skills, ~/.gemini/config/skills, ~/.config/opencode/skills). Or install through the finder skill, which does both.
Badge for your README
The badge for sickn33/agentic-awesome-skills shows the score of its lowest-scoring skill.
[](https://zalavijaysinh.github.io/skilltrust/repos/sickn33/agentic-awesome-skills/)
Score history
| Date | Version | Score |
|---|
| 2026-11-11 | 00f9c40 | 59 |
| 2026-11-08 | 828257c | 59 |
| 2026-11-07 | 828257c | 59 |