← All skills
huggingface-spaces
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 huggingface/skills (GitHub) · coding · 11,321 stars · scanned at commit 25910b3
Build, deploy, and maintain applications on Hugging Face Spaces — Gradio / Docker / Static SDKs, ZeroGPU and dedicated hardware, model loading, debugging, buckets, inference providers, community grants. Use whenever the user asks to create or host an app on Hugging Face, port code onto ZeroGPU, fix a Space that won't build or run, or otherwise work with `hf spaces …`, `@spaces.GPU`, Space README frontmatter, or the `spaces` Python package.
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 20 places.
Show 17 more
- skills/huggingface-spaces/references/3d-cuda-extensions.md, line 39
subprocess.run(shlex.split("pip install custom_rasterizer-0.1-cp310-cp310-linux_x86_64.whl"), check=True) - skills/huggingface-spaces/references/3d-cuda-extensions.md, line 59
shlex.split("pip install --no-build-isolation ./texture_baker"), - skills/huggingface-spaces/references/3d-cuda-extensions.md, line 65
*Inside a one-shot `@spaces.GPU(duration=600)` setup function called at module import* (`trellis-community/TRELLIS`, `dylanebert/LGM-mini` — the latter for the - skills/huggingface-spaces/references/3d-cuda-extensions.md, line 67
JIT costs: slower cold starts, a burned GPU allocation (second variant), and breakage when the runtime image moves. Prefer wheels; when you do JIT-build, log cl - skills/huggingface-spaces/references/3d-cuda-extensions.md, line 109
6. App RUNNING but a feature is missing (e.g. texture button gone) → a startup `pip install`/compile failed silently; read the run log from the top. - skills/huggingface-spaces/references/3d-generation.md, line 95
- Compiling CUDA extensions via `requirements.txt` (no nvcc at build time), or `pip install`-ing PyPI lookalike packages for the model libraries. - skills/huggingface-spaces/references/3d-gsplat.md, line 23
**Shape B — a real rasterizer (only for server-rendered orbit videos).** If the code uses the upstream Inria API, take the prebuilt wheel from `requirements.md` - skills/huggingface-spaces/references/3d-models.md, line 14
There is no usable pip package (`trellis-3d` on PyPI ships only the pure-Python tree without the CUDA extensions or real dependency set; nothing exists for TREL - skills/huggingface-spaces/references/3d-models.md, line 73
- `pip install --no-build-isolation ./extensions/nvdiffrast` (vendored source), then the mip-splatting `diff-gaussian-rasterization` from a shallow git clone - skills/huggingface-spaces/references/3d-models.md, line 126
shlex.split("pip install --no-build-isolation git+https://github.com/tatsy/torchmcubes.git"), - skills/huggingface-spaces/references/3d-models.md, line 156
"pip install -vv --no-build-isolation ./texture_baker ./uv_unwrapper" - skills/huggingface-spaces/references/3d-models.md, line 187
subprocess.run(shlex.split("pip install custom_rasterizer-0.1-cp310-cp310-linux_x86_64.whl"), check=True) - skills/huggingface-spaces/references/debugging.md, line 199
It's a normal container at `/home/user/app/`. You can edit files, `pip install`, run repros, call `@spaces.GPU`-decorated functions interactively (they get a re - skills/huggingface-spaces/references/requirements.md, line 56
Verification after `pip install` or `uv lock --upgrade`: - skills/huggingface-spaces/references/requirements.md, line 68
This is the **first** thing to reach for when a CUDA-extension package has no upstream wheel matching the ZeroGPU torch / CUDA / cxx11-abi cell. Prefer it over - skills/huggingface-spaces/references/requirements.md, line 106
| `pytorch3d` | a runtime `pip install git+…pytorch3d.git` inside `@spaces.GPU` | **0.7.9**. Needs `numpy`, `iopath`, `fvcore` listed. No torch pin in metadata; - skills/huggingface-spaces/references/requirements.md, line 109
| `torchmcubes` | a runtime `pip install git+…torchmcubes.git` | **0.1.0**. **sm_120 only** (no fatbin for older archs). Works on ZeroGPU / Blackwell; not porta
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 17 places.
Show 14 more
- skills/huggingface-spaces/SKILL.md, line 185
hf spaces logs <ns>/<name> --tail 200 - skills/huggingface-spaces/SKILL.md, line 212
hf buckets create <ns>/<bucket-name> # --private optional - skills/huggingface-spaces/references/3d-cuda-extensions.md, line 22
bash -c 'export TORCH_CUDA_ARCH_LIST="12.0+PTX" && \ - skills/huggingface-spaces/references/3d-cuda-extensions.md, line 39
subprocess.run(shlex.split("pip install custom_rasterizer-0.1-cp310-cp310-linux_x86_64.whl"), check=True) - skills/huggingface-spaces/references/3d-cuda-extensions.md, line 49
os.system("cd /home/user/app/hy3dpaint/DifferentiableRenderer && bash compile_mesh_painter.sh") - skills/huggingface-spaces/references/3d-cuda-extensions.md, line 58
subprocess.run( - skills/huggingface-spaces/references/3d-cuda-extensions.md, line 67
JIT costs: slower cold starts, a burned GPU allocation (second variant), and breakage when the runtime image moves. Prefer wheels; when you do JIT-build, log cl - skills/huggingface-spaces/references/3d-models.md, line 125
subprocess.run( - skills/huggingface-spaces/references/3d-models.md, line 154
os.system( - skills/huggingface-spaces/references/3d-models.md, line 187
subprocess.run(shlex.split("pip install custom_rasterizer-0.1-cp310-cp310-linux_x86_64.whl"), check=True) - skills/huggingface-spaces/references/3d-models.md, line 189
os.system("cd /home/user/app/hy3dpaint/DifferentiableRenderer && bash compile_mesh_painter.sh") - skills/huggingface-spaces/references/known-errors.md, line 89
subprocess.run([sys.executable, "-m", "pip", "install", "--no-deps", - skills/huggingface-spaces/references/requirements.md, line 42
subprocess.run([sys.executable, "-m", "pip", "install", "--no-deps", - skills/huggingface-spaces/references/requirements.md, line 128
**Do not** install these from `@spaces.GPU` startup. A `subprocess.check_call` pip-install at first GPU acquire is strictly worse than the wheel URL — slower co
Sends data to these websites: huggingface.co, pypi.org
This is disclosed by the author in the skill's SKILL.md or README.
Found in 7 places.
Show 4 more
Reads your saved credentials, such as SSH keys or .env files
The author does not mention this.
Found in 2 places.
Safer alternatives
Low-risk skills in the same category (coding) that look related. Check that one does what you need.
Install
Paste this into your terminal. It downloads exactly the version we scanned, not whatever is newest.
if [ -e ~/.agents/skills/'huggingface-spaces' ] || [ -L ~/.agents/skills/'huggingface-spaces' ]; then echo 'A skill folder named huggingface-spaces already exists — not replacing it' >&2; false; else d=$(mktemp -d) && git clone --quiet --filter=blob:none --no-checkout 'https://github.com/huggingface/skills' "$d" && git -C "$d" checkout --quiet '25910b3d0ba31a2ba102f901cdc9e008c278a69c' -- 'skills/huggingface-spaces' && mkdir -p ~/.agents/skills && cp -R "$d"/'skills/huggingface-spaces' ~/.agents/skills/'huggingface-spaces' && 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/'huggingface-spaces' ~/.claude/skills/'huggingface-spaces' (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 huggingface/skills shows the score of its lowest-scoring skill.
[](https://zalavijaysinh.github.io/skilltrust/repos/huggingface/skills/)
Score history
| Date | Version | Score |
|---|
| 2026-11-08 | 25910b3 | 59 |
| 2026-11-06 | 25910b3 | 69 |