GitHub

GitHub is split into three areas: issues, pull requests and repos with code, files, branches and releases. The assistant reads in them and can create or comment on issues and pull requests when you ask it to.

• “Which issues have been open for more than a month?”
• “Summarize what this pull request changes.”
• “Which releases were there in this repository since July?”
• “Create an issue: sign-in fails with long names.”

1. 01In anymize open Apps, then Connectors, and search for GitHub.
2. 02Click Connect and sign in to GitHub once.
3. 03Read the permission GitHub shows you and confirm it. For an organization their consent may also be needed.
4. 04Then choose which areas the assistant may use: issues, pull requests, repos.

The permission GitHub shows you also covers private repositories. That is needed so skills can be imported from a private repository, and it is exactly the point you should look at before you confirm.

Two things limit that in everyday use. First you see in GitHub's permission exactly what is asked, and you can decline. Second you then decide in anymize which areas are on at all. Anyone who only wants to analyse issues leaves repos off.

Names, email addresses, customer names and credentials that appear in issues, comments or files are replaced by placeholders before the content goes to an AI model. A bug report stays understandable; the customer name in it does not leave your workspace in clear text.