started · updated
CoderBuds open-sources AI code detection rules
CoderBuds has open-sourced its AI code detection rules to provide engineering teams with transparency regarding AI-generated pull requests and commits. This move comes as GitHub reports that over 46% of code on its platform is now AI-assisted.
The detection framework is designed to be language-agnostic, supporting PHP, Python, Node.js, Ruby, and Go. It utilizes specific markers, footers, and patterns to identify various AI coding assistants, including Anthropic’s Claude Code, Microsoft’s GitHub Copilot, Cursor, Devin, and OpenAI’s Codex.
In practical application, tools like Claude Code function as command-line agents that can read codebases, edit files, and execute engineering tasks within a user's working directory. Users can access these tools through native installers or package managers like npm, with usage models ranging from flat-rate subscriptions to pay-as-you-go API billing.
Entities
Anthropic · Claude Code · CoderBuds · GitHub · Microsoft
Claims
What the coverage asserts, and how many sources carry each claim.
- [○ 1 SOURCE] The detection rules can identify tools including Claude Code, GitHub Copilot, and Cursor. packagist.org
- [○ 1 SOURCE] More than 46% of code on GitHub is currently AI-assisted. packagist.org
- [○ 1 SOURCE] CoderBuds has open-sourced its AI code detection rules. packagist.org