Simon Willison released commit-rewriter 0.1, a Python web app that rewrites commit messages to remove coding-agent artifacts and private issue references. He built it to prepare the Datasette September security releases for publication, whose initial commits contained agent cruft and private repository issue IDs. It runs via uvx and creates a timestamped backup branch before rewriting, so edits can be reverted.
A developer open-sourced subcommit, an MIT-licensed CLI that lets agents commit only their changes while preserving unrelated staged, unstaged, or untracked changes. It targets the pain of working with multiple concurrent agents and user edits in one repo. The author has used it for a few months before releasing it.
A developer open-sourced subcommit, an MIT-licensed CLI that lets agents commit only their changes while preserving unrelated staged, unstaged, or untracked changes. It targets the pain of working with multiple concurrent agents and user edits in one repo. The author has used it for a few months before releasing it.
Auracommit is a CLI tool that generates Git commit messages using AI, processing diffs file by file to minimize token usage. It supports multiple AI providers and helps developers create detailed commits without exceeding API limits.
Auracommit is a CLI tool that generates Git commit messages using AI, processing diffs file by file to minimize token usage. It supports multiple AI providers and helps developers create detailed commits without exceeding API limits.
Gitwhy is a tool that connects AI coding sessions to Git commits, ensuring that the context of AI-assisted changes is preserved before commits are deleted. It solves the problem of losing valuable context from AI pair programming sessions, benefiting developers who use AI tools and want to maintain a clear history of their work.
Gitwhy is a tool that connects AI coding sessions to Git commits, ensuring that the context of AI-assisted changes is preserved before commits are deleted. It solves the problem of losing valuable context from AI pair programming sessions, benefiting developers who use AI tools and want to maintain a clear history of their work.
Researchers propose CommitLLM, a three-stage pipeline that generates concise, Conventional Commits-compliant git commit messages from code diffs. The system uses QLoRA fine-tuning of Mistral-7B-Instruct-v0.2 on the CommitPackFT dataset, constrained decoding for brevity, and post-processing to enforce format. The work addresses the common problem of uninformative commit messages in version control.
Researchers propose CommitLLM, a three-stage pipeline that generates concise, Conventional Commits-compliant git commit messages from code diffs. The system uses QLoRA fine-tuning of Mistral-7B-Instruct-v0.2 on the CommitPackFT dataset, constrained decoding for brevity, and post-processing to enforce format. The work addresses the common problem of uninformative commit messages in version control.
Simon Willison published a GitHub code-frequency chart for his Datasette project, showing a dramatic spike in code activity coinciding with the release of advanced AI coding models like Opus 4.8, GPT-5.5, Fable 5, and GPT-5.6 Sol. The chart illustrates the tangible impact of AI coding agents on developer output.
Simon Willison published a GitHub code-frequency chart for his Datasette project, showing a dramatic spike in code activity coinciding with the release of advanced AI coding models like Opus 4.8, GPT-5.5, Fable 5, and GPT-5.6 Sol. The chart illustrates the tangible impact of AI coding agents on developer output.