Guide · Agent skill
How to use the SageMaker Python Environment Setup agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
This guide explains what the hf-cloud-python-env-setup agent skill does, how to install it for Claude, and how to ask your agent to use it. It is for developers who prepare SageMaker deployments, training jobs or other AWS automation and keep running into Python version or dependency errors. By the end you will know how to get the agent to build an isolated .venv with a supported Python and current boto3 and awscli before it runs any code.
SageMaker Python Environment Setuphuggingface/skills · Apache-2.0 Download the skill Quick start
- 1Download the ZIP. Download the skill ZIP from the skill's page and confirm it contains SKILL.md, requirements.txt and the scripts folder.
- 2Install the skill. Unzip it into ~/.claude/skills/ for personal use or into a project's .claude/skills/ folder, or upload it under Settings, Capabilities, Skills in the Claude apps.
- 3Confirm Python is available. Make sure Python 3 is installed; 3.10 to 3.12 is recommended. uv is optional.
- 4Ask for the environment. Make a request such as set up the environment for this SageMaker deployment, so the agent loads the skill.
- 5Let the script build .venv. The agent runs scripts/setup_env.py, which creates .venv and installs boto3 and awscli from requirements.txt.
- 6Verify versions. Run scripts/check_versions.py with the env's Python to confirm boto3, botocore and awscli are installed.
- 7Run code with the env's interpreter. Call .venv/bin/python (or .venv\Scripts\python.exe on Windows) for your deployment or AWS scripts.
What an agent skill is and what this one makes the agent do
An agent skill is a folder of instructions and optional scripts that an AI agent loads when a task matches the skill's description. This skill is the environment-setup step for SageMaker and AWS work. Its SKILL.md says that most SageMaker deployment failures that look like AWS problems are really Python environment problems, such as the wrong Python version, broken dependency resolution or an outdated SDK.
Once the skill applies, the agent sets up a separate isolated Python environment before running code that talks to AWS or SageMaker, instead of using the computer's main Python. It creates a .venv folder using a supported Python version (3.10 to 3.12), installs the latest boto3 and awscli from the bundled requirements file, and checks installed package versions.
- Never uses the system Python or runs pip install into it.
- Creates .venv with Python 3.12 by default, or another supported version you specify.
- Uses uv if it is available and falls back to the standard library venv module otherwise.
- Installs the latest boto3 and awscli rather than pinned versions.
- Re-running the setup script is safe because it is idempotent.
Install the skill (ZIP download or Claude app upload)
Download the skill ZIP from the skill's page in this directory, then choose one of two routes. For a personal setup, unzip the folder into ~/.claude/skills/. For a single project, unzip it into the project's .claude/skills/ folder so the skill applies only there.
If you use the Claude apps instead, upload the skill under Settings, then Capabilities, then Skills. Either way the folder should contain SKILL.md, requirements.txt and the scripts folder with setup_env.py and check_versions.py.
- Personal: unzip into ~/.claude/skills/
- Project only: unzip into .claude/skills/ inside the project
- Claude apps: Settings → Capabilities → Skills, then upload
Ask the agent to use it (trigger phrases)
You usually do not need to name the skill. The skill's description tells the agent to use it whenever Python code will run for a SageMaker deployment, training job or AWS automation. That includes before running pip install, before invoking boto3, when creating or activating a virtualenv, or when you ask it to set up the environment.
Plain requests work well. Mention SageMaker or AWS so the agent knows which rules apply.
- 'Set up the environment for this SageMaker deployment.'
- 'Create a virtualenv for the AWS automation and install what it needs.'
- 'Before you pip install anything for the training job, set up a clean Python environment.'
- 'Check which boto3 version is installed in the project environment.'
Run the setup and check the versions
The fastest path is the bundled setup script, which is written in Python so it runs the same on Windows, macOS and Linux. It creates .venv/ with Python 3.12 unless you give a different version, refuses unsupported versions, installs from requirements.txt and prints the correct interpreter path for your OS.
To verify the result, run the version checker. It reports boto3, botocore and awscli, and you can pass other package names to check them too. Because the skill is idempotent, running setup again on an existing environment is safe.
python3 scripts/setup_env.py
python3 scripts/setup_env.py .venv 3.11
.venv/bin/python scripts/check_versions.py
.venv/bin/python scripts/check_versions.py transformers huggingface_hubRun your code with the environment's Python
After setup, the skill has the agent call the environment's Python directly rather than activating the venv. This works the same in scripts, interactive shells and agent tool calls. On Windows, use the Scripts path instead of bin.
The source's manual equivalents are shown below. Use them only if you cannot run the script.
- Windows path: .venv\Scripts\python.exe deploy.py
- If a package seems missing, install it with the same interpreter that runs your script.
uv venv --python 3.12 .venv
uv pip install --python .venv/bin/python --upgrade boto3 awscli
python3.12 -m venv .venv
.venv/bin/python -m pip install --upgrade pip boto3 awscliAdd extras and understand what the skill does not do
The default requirements.txt covers SageMaker orchestration only. If a deployment needs more, such as huggingface_hub for model inspection or transformers for tokenizer validation, add them to a deployment-specific requirements file in your project. The skill does not pin versions unless you have a documented reason.
The skill prepares the environment and does not deploy models. The deploy scripts it refers to (deploy.py, deploy_async.py, teardown.py) are not included in these files, so you need your own copies to run deployments. The SageMaker Python SDK is optional and is not needed by the bundled scripts.
- Pinned boto3 or awscli versions must be added by you manually.
- Non-AWS Python projects are outside this skill's rules.
- AWS credentials are needed later for real AWS calls, not for setup.
What people use it for
- Preparing a SageMaker deployment when earlier attempts failed with pip resolution or module-not-found errors.
- Setting up a reproducible environment for AWS automation shared across a team.
- Making sure an agent does not install packages into the system Python while working on an AWS task.
- Checking that boto3 and awscli are current before calling a newer AWS API.
- Adding huggingface_hub or transformers to a project-specific requirements file alongside the base requirements.
Frequently asked questions
Is the SageMaker Python Environment Setup skill free to use?
The source lists the license as Apache-2.0, which is the license for the skill's files. The source does not discuss pricing, so check the skill's page for any other terms. Running the setup itself does not need AWS credentials.
What do I need before I install this skill?
You need Python 3 on your machine, with 3.10 to 3.12 recommended. uv is optional because the script falls back to the standard library venv module. The setup needs network access to install boto3 and awscli, and AWS credentials are only needed later for real AWS calls.
Where do I put the skill so Claude can find it?
For personal use, unzip the skill folder into ~/.claude/skills/. For one project, put it in that project's .claude/skills/ folder. You can also upload it in the Claude apps under Settings, then Capabilities, then Skills.
Why does the skill use Python 3.12 and not the newest Python?
The skill's rules say to use 3.10, 3.11 or 3.12 and to avoid 3.13 and later, because ML libraries can lag on wheel availability and dependency resolution can break in confusing ways. The setup script defaults to 3.12 and refuses unsupported versions.
My pip install fails with resolution errors. What should I do?
The skill's common pitfalls say this is usually caused by Python 3.13 or later, or by installing into a polluted system Python. Delete .venv and re-run python3 scripts/setup_env.py .venv 3.12, or simply re-run the script, since it recreates the environment when the version does not match.
How does this compare with the SageMaker Python SDK?
The bundled scripts use boto3 directly and do not need the SageMaker Python SDK. The SKILL.md says the SDK is a valid alternative if you or your project prefer it, and you can install it into the same isolated environment. Installing it does not replace the requirements install.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use this skill if you prepare SageMaker deployments or AWS automation and want an agent to build a clean, isolated Python environment before running code, especially if you keep hitting version or dependency errors. Pick something else if you need pinned boto3 or awscli versions, need the skill to deploy models for you, or work on non-AWS Python projects.
Download the skill