Python Requirements.txt vs Imports Checker

Find dependency drift before deployment does

Python projects accumulate dependencies over time. Some stay in requirements.txt long after the code stops using them, while other packages are imported in source code but never declared at all.

Python Requirements.txt vs Imports Checker scans the actual project locally and creates a clean dependency discrepancy report.

• Recursively scans Python files using the AST
• Ignores virtual environments, build folders, caches and site-packages
• Excludes Python standard-library modules and local project packages
• Detects imports that appear to be missing from requirements.txt
• Flags declared packages with no observed static import
• Handles common package/import mismatches such as Pillow → PIL, PyYAML → yaml, scikit-learn → sklearn and python-dotenv → dotenv
• Recognizes literal dynamic imports when possible
• Reports ambiguous dependency entries instead of guessing
• Generates Markdown and JSON reports

Why use the Skill instead of pasting files into chat?

The Skill recursively inspects the real local repository, parses Python imports structurally, excludes local and standard-library modules, and applies the same dependency rules every time.

AST Scan → Import Map → Requirements Parse → Dependency Diff → Reviewable Report

Free by design

This free Skill detects dependency drift. Automatic dependency migration, lockfile regeneration, Poetry/Pipenv/uv conversion, compatibility resolution and security remediation remain outside this lightweight checker.

## What it does - **AST-based scanning** recursively analyzes `.py` files to find all top-level and literal dynamic imports. - **Dependency mapping** resolves discrepancies between distribution names and import aliases like Pillow/PIL or PyYAML/yaml. - **Noise reduction** automatically filters out standard library modules, local relative imports, and virtual environment directories. - **Discrepancy classification** flags imports missing from your requirements and declared packages that appear unused. - **Report generation** produces both Markdown and JSON summaries for manual review or CI/CD integration. ## How it works 1. Run the `import_scanner.py` script pointing to your project root and `requirements.txt`. 2. The agent executes an AST scan, excluding `venv`, `node_modules`, and standard library modules. 3. The tool compares findings against the provided requirements file using the internal alias map. 4. Review the generated `requirements_imports_report.md` for prioritized action items. ## Frameworks & tools This skill uses the Python `ast` module for static analysis and targets standard `requirements.txt` files. It includes specific support for common libraries like scikit-learn, OpenCV, and SQLAlchemy. ## Why this beats prompting it yourself Generic prompts often fail to recognize that `scikit-learn` is imported as `sklearn`, leading to false positives. This skill uses a curated mapping file and AST parsing to ensure it doesn't just grep strings, providing a much higher signal-to-noise ratio. ## Use cases - Identify missing dependencies before a fresh deployment or CI build. - Prune unused packages from a legacy monolith to reduce container image size. - Audit a project after a large refactor to sync declarations with source. - Standardize dependency management across multiple microservices. ## Known limitations Static analysis cannot guarantee detection of all dynamic imports, plugins, or CLI-only tools. It does not support Poetry or Conda lockfiles directly.