What probably matters to youSection of the release
Unclear It is not stated what happens when the cap is reached, or what the default cap is.
What
A runner is the process that prepares a repository for a session that does not run on your own machine. It uses git fetch to download the repository's history. A new environment variable, CLAUDE_RUNNER_FETCH_SERVER_PROGRESS_CAP_MS, sets a time cap on the server-side part of that fetch. An environment variable is a named value you set before starting a program. It accepts:
a positive whole number of milliseconds, kept between two minutes and thirty minutes
Any other value is ignored and the default is used. Claude Code also now reads the "remote: Enumerating objects: N" lines the server prints while preparing a fetch.
Why
This matters for cloud and self-hosted runners cloning slow or very large repositories. You can now change or remove the cap if a fetch needs more time.
Something disagreesSomething we can check disagrees with this entry, or the writer said they could not settle it.
The writer flagged doubtIt is not stated what happens when the cap is reached, or what the default cap is.
Anthropic's release notes agreeSelf-hosted runner: Fixed git fetches being cut off while the git server was still reporting progress on a large repository…
The name it cites is new in this buildNew in this build: CLAUDE_RUNNER_FETCH_SERVER_PROGRESS_CAP_MS
