Back to Cookbook

Env Doctor

Fix Python dependency hell without losing your afternoon

Diagnose and fix conda/pip conflicts, generate reproducible environment files, and containerize analytics projects. No more "it works on my machine" or spending an entire day getting dependencies to resolve.

CommunitySubmitted by CommunityWork1 min

PROMPT

Create a skill called "Env Doctor". Help me diagnose and fix Python environment issues. When I paste an error message (dependency conflicts, ImportError, version mismatches, conda solver failures), identify the root cause and provide a specific fix. When I point you at a Python project, analyze all imports and generate the minimal environment file (requirements.txt, environment.yml, or pyproject.toml) needed to run it. Detect conda/pip mixing issues (packages installed by both managers). If an environment is beyond repair, generate a Dockerfile that creates a clean, reproducible environment. Support conda, pip, poetry, and uv. For data science projects, know the common conflict patterns (e.g., numpy binary compatibility, tensorflow + CUDA version matrix, scikit-learn + scipy version coupling).

How It Works

Analytics teams lose entire days to Python environment issues — conda solver

hanging for hours, pip and conda fighting over numpy versions, a colleague's

notebook that only runs on their machine. This skill untangles the mess.

What You Get

  • Dependency conflict diagnosis from error messages
  • Auto-generated environment files (requirements.txt, environment.yml, pyproject.toml)
  • Conda/pip conflict detection and resolution
  • Environment recreation from a script's import statements
  • Docker containerization for reproducible execution
  • Migration between conda, pip, poetry, and uv

Setup Steps

  1. Ask your Claw to create an "Env Doctor" skill with the prompt below
  2. Paste the error message, point it at your project, or describe the issue
  3. Get back a diagnosis and fix

Tips

  • Prevention beats cure: run the audit before sharing code with teammates
  • The Docker containerization is the nuclear option when nothing else works
  • uv is often the fastest fix for pip dependency resolution issues
  • For data science stacks (numpy, scipy, scikit-learn, pandas), the skill knows the common version conflicts
Tags:#python#environment#devops#productivity