AI & Agents

What’s New in Claude Code 2.1.325: A zstd Decode Fix and a Lot of Sharpening

Leave a comment

Claude Code 2.1.325 is out, and the release notes are real, not a stub. The first item is a security fix in how zstd-compressed responses are decoded. Anthropic doesn’t describe an attack in the notes, and I’m not going to invent one. If you run Claude Code, this is a good reason to pick up the release.

The rest is a fairly large polish pass. There’s a one-time VS Code tip about re-running a notebook cell once Claude has touched a Jupyter notebook, and a clearer next step when an organization policy blocks a request, pointing you to /rewind or /clear. Fixes include /security-review and inline git commands in unusual git layouts, subagents staying on their own model instead of getting pulled onto the parent plan-mode model, sessions on some third-party gateways compacting instead of erroring out on every turn, recent turns going missing after an unclean exit, and a few UI and plan-saving edge cases.

Review before you install: read the release notes yourself, back up your config, and try new versions somewhere you can afford to break first.

My take

A security fix in response decoding is the item I’d actually upgrade for, even if the rest is “just” a lot of sharp edges getting filed down. Read the notes, then move. I wrote about 2.1.323’s experimental Code Mode two days ago; this release is a different, quieter kind of necessary.

God Bless, and May the Source Be with You!

References & Further Reading

Subscribe (RSS) Follow AI & Agents via RSS

Leave a note

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.