Claude Code's week: Sonnet 5.5 by default, Mods and a side agent
Sonnet 5.5 became the default Sonnet, Mods let plugins change how Claude Code behaves, and a built-in side agent flags what you miss.
By the AyCode team
Researched and written with AI, edited and approved by Niccolo Bodenstedt.
Seven releases in eight days. Claude Code's changelog from 29 September to today reads like a team that does not sleep, and three changes stand out for anyone who codes with it every day.
Sonnet 5.5 is the new default Sonnet
Version 2.1.286 (30 September) added Claude Sonnet 5.5 (claude-sonnet-5-5) and made it the default Sonnet, with a 1M token context window at $2 per million input and $10 per million output tokens. One release earlier, 2.1.285 had already moved interactive sessions to the 1M window for Opus 4.7 and later, Sonnet 5 and later, and Fable.
Mods: plugins that change behavior, not just add tools
Version 2.1.290 (5 October) introduced Claude Mods. Plugins could already add commands, agents and hooks; Mods let them modify deeper behavior and are shared through the plugin directory. 2.1.289 had given mods $.ui.selection(), which returns the text you selected in fullscreen mode, and 2.1.290 adds more data to their hook results.
The first built-in one is called "You should know": a side agent that watches the session and flags things you or Claude might miss. You turn it on with /plugin enable cc-plugin-you-should-know@builtin. The changelog says it is available in first-party sessions with telemetry on.
The small things that save your afternoon
/code-reviewtakes--max-findings <n>orall(2.1.289).- Mid-response API timeouts no longer throw away the answer: non-interactive sessions continue from the partial response (2.1.289).
- A prompt cleared with Ctrl+C can be recovered (2.1.288).
- WebFetch no longer cuts pages silently after 100,000 characters; it says so and offers an offset (2.1.290).
- Scheduled tasks no longer stop running after a conversation is compacted (2.1.290).
Our take
Mods are the bigger story than another model bump. A side agent that reviews while you work is the kind of thing teams used to build on their own, badly. The catch is the telemetry requirement for the built-in one, which some companies will not accept.
In AyCode
In AyCode, Sonnet 5.5 is already in the model menu. The usage view shows what each chat costs in tokens, so a bigger context window does not turn into a surprise bill.
Sources
Where every claim in this post comes from. Posts on X are linked, never copied.
- AnthropicClaude Code changelog