Follow Discord
Sweep 25 Sep 2026 · 19:33Z Build v2.1.283 504 read Stable v2.1.274 Latest v2.1.283 Next v2.1.283 Feeds RSS JSON llms.txt llms-full.txt Unofficial

Claude Code v2.1.283 ·

Directory sync treats a full store as non-blocking, and a full lane now counts as store_full

When directory sync's store is full, remote commands now run anyway instead of failing, and a full link lane is reported as store_full

Group of 3 You'll notice Improvements
JSON All of v2.1.283
You'll noticeTier: how much it should matter to you
2Useful: my rating, 1 to 5
1Signal: worth watching, 1 to 5
Cloud SyncArea: what it touches
ImprovementsKind: in v2.1.283,
ImprovementsSection of the release

What

Directory sync sends your session's file changes to another machine before a command runs there. When the store holding those changes is full (store_full):

  • The failure is still recorded, but the exchange with the other machine carries on. Before, any failed push stopped at that point.
  • The remote command now just runs, instead of failing with a message and a suggested fix.
  • store_full was added to the file-sync failure reasons.

In the git-based sync worker, a lane_full answer while detached now marks the store as full. Before, it went through the generic not-sent path and was counted as a failed upload. Bundles that are too large no longer take a separate early exit in that mode. A record rebuilt from an earlier stored row now carries its real unshipped flag instead of always false.

Why

A full sync store no longer blocks commands on the other machine. The catch is that such a command may run on files that are missing your latest changes, and you get no note saying so.

How sure we are
Something disagreesSomething we can check disagrees with this entry, or the writer said they could not settle it.
The writer flagged doubtWhat this means for the command itself or for what a user sees is not clear.

See this entry in the whole of v2.1.283 →

Feedback