Follow Discord
Sweep 29 Sep 2026 · 18:10Z Build v2.1.285 506 read Stable v2.1.280 Latest v2.1.285 Next v2.1.285 Feeds RSS JSON llms.txt llms-full.txt Unofficial

Claude Code v2.1.285 ·

Auto mode always lets Claude Code's own tool calls skip the safety check

In auto mode, tool calls that Claude Code generates itself now always skip the classifier, with no remote switch to turn this off

You'll notice Improvements
JSON All of v2.1.285
You'll noticeTier: how much it should matter to you
2Useful: my rating, 1 to 5
2Signal: worth watching, 1 to 5
Auto ModeArea: what it touches
ImprovementsKind: in v2.1.285,
ImprovementsSection of the release
What

In auto mode, a classifier (a model that judges whether an action is safe) usually checks each tool call Claude makes. Tool calls that Claude Code generates itself already skipped that check. That shortcut used to sit behind a remote switch that was on by default. The switch has been removed, so the shortcut now always applies.

Why

The shortcut can no longer be turned off from the server. Tool calls that Claude Code generates itself will skip the classifier in auto mode in every case where the shortcut applies.

Read from
Feature flag
tengu_structured_creek Not enough to say

Nothing here resolved what this flag was doing on this version, so nothing here should be read as on or off.

This account: no value returned · anonymous baseline: no value returned · compiled default in v2.1.285: not a boolean we can read

These values were read against a different version of Claude Code, so treat them as the nearest reading available instead of one taken on this release.

Read once, for one account on one subscription tier, against v2.1.285. It isn't a statement about your account. What a flag value here can and cannot tell you

See this entry in the whole of v2.1.285 →

Feedback