Git Gibberish Checker

The Git Gibberish Checker analyzes commit messages to identify signals of low semantic quality and noisy wording. It tokenizes each message, computes a gibberish_score using entropy and dictionary-based heuristics, and flags entries that fail a configurable threshold. The tool provides per-message feedback and suggested edits to reinforce concise, descriptive messages. It is designed for integration into code review workflows and CI pipelines, offering machine-friendly outputs for automation and human-friendly explanations for reviewers.

Conceptually, it separates signals from noise: high-entropy tokens, leetspeak, all-caps segments, or semantically hollow phrases increase the score, while clear, action-oriented phrases reduce it. It supports optional localization and per-repo tuning to reflect project conventions. The unique value lies in actionable guidance: beyond a simple pass/fail, it returns concrete suggestions that align with semantic commit practices and helps teams improve traceability across releases. Typical use cases include pre-commit checks, CI checks, and onboarding new contributors who must adhere to commit standards.

Key features include per-message scoring, configurable thresholds, optional diff analysis, and structured JSON outputs suitable for dashboards or automated reviews. By focusing on message clarity rather than syntax alone, it accelerates reviews, reduces misunderstandings, and promotes a consistent commit history across teams.