plankton-code-quality

Automated code formatting and linting during write-time is achieved by using Plankton post-tool hooks to intercept Claude edits and run linters like ruff, biome, and hadolint on each modification. This enforces code quality before changes enter the main task.

Yes, automated linting and repairs support multiple languages including Python, TypeScript, Shell, YAML, Markdown, Dockerfile, JSON, and TOML. A multi-model routing system assigns repair tasks to Haiku, Sonnet, or Opus based on violation severity.

To enforce code quality with Plankton hooks, you need to clone the Plankton project, install language linters such as ruff, biome, markdownlint, and hadolint, and configure pre-tool and stop hooks for protection during your Claude-driven workflow.

Automated code repair uses Claude subprocesses to fix violations when in-process checks miss issues. The multi-model routing system drives these targeted repairs by assigning tasks to different model tiers based on the severity of the formatting or linting violation.

Yes, write-time code quality enforcement works with Dockerfile and YAML formats, alongside JSON, TOML, Shell, and Markdown. Plankton hooks trigger formatting and linting tools like hadolint across these languages automatically on each edit.

The best way to prevent style and correctness issues is on-the-fly enforcement using post-tool-use hooks that format and lint code as you edit. This write-time approach catches violations during generation rather than post-commit.