This guide is for people who don't write code. You'll use an AI coding assistant — Claude Code, GitHub Copilot, or OpenAI Codex — to do all the technical work for you. Your job is just to:
- Open your AI assistant.
- Copy the prompt below and paste it in.
- Answer a couple of simple questions when the assistant asks.
The AI will create a fresh project folder, install the COMcheck API library, and set up the "Skill" that teaches the AI how to use the library correctly.
You need two things:
- Go to https://comcheck.energycode.pnl.gov and log in (or register).
- Click your username in the left-side menu.
- Choose Settings → Developer Setting.
- Click Generate, then copy the token immediately.
⚠️ The token is shown only once. Paste it somewhere safe (a password manager or a note) before leaving the page. If you lose it, just generate a new one.
- Claude Code — in your terminal, or the VS Code / JetBrains extension.
- GitHub Copilot — the Chat/Agent panel in VS Code.
- OpenAI Codex — the Codex CLI or extension.
Have the AI assistant open and ready to chat. Then continue below.
Select everything inside the box and paste it into your AI assistant's chat.
Please set up a brand-new Python project for me that uses the COMcheck API library. I am not a programmer, so do every step for me and explain what's happening in plain language. Specifically:
- Create a new project folder called
comcheck-projectand open it.- Make sure Python 3.12 or newer is available. If a tool called
uvis installed, use it; otherwise use plainpipand a virtual environment. Pick whichever works and tell me what you chose.- Install the library from PyPI: the package name is
comcheck-api.- After it's installed, run the command
comcheck-api install-skillinside the project folder. This installs the bundled AI "Skill" so you (the assistant) know how to use the library correctly. Confirm that it created a.claude/skills/comcheck-api/folder (and/or.agents/skills/comcheck-api/).- Create a file named
.envin the project folder with this line, and pause to let me paste in my real key:COM_API_KEY=<I will paste my key here>- Add
.envto a.gitignorefile so my key never gets shared.- Write a small test script called
check_setup.pythat loads the key from.env, creates aCOMcheckClient, and prints a friendly message confirming the library is installed and the key was read. Run it and show me the output.Go one step at a time. Stop and ask me whenever you need my API key or a decision from me.
When the assistant pauses at the .env step, paste the key you copied
earlier so the line looks like:
COM_API_KEY=abc123yourrealtokenhere
Save the file (the assistant can do this for you).
When the assistant runs check_setup.py, you should see a message like
"COMcheck API is installed and your key was loaded." If you see an error
instead, paste the error back to the assistant and ask it to fix it — that's
exactly what these tools are good at.
A folder named comcheck-project containing:
- The
comcheck-apilibrary, installed and ready. - A Skill folder (
.claude/skills/comcheck-api/and/or.agents/skills/comcheck-api/) that makes your AI assistant an expert on this library. - A
.envfile holding your API key (kept private by.gitignore). - A working
check_setup.pytest script.
From here you can ask your AI assistant things like "Run a COMcheck compliance simulation for a small office building" and it will use the installed Skill to write the right code for you.
The whole setup boils down to these commands, which the AI runs for you:
# Install the library
pip install comcheck-api # or: uv add comcheck-api
# Install the AI Skill into this project
comcheck-api install-skill # both Claude & Codex
# comcheck-api install-skill --claude # Claude Code only
# comcheck-api install-skill --codex # OpenAI Codex only- Requires Python 3.12+.
- Full docs: https://pnnl.github.io/comcheckweb-api-python/
- Get an API key: https://comcheck.energycode.pnl.gov