Team memory sync will gain a manifest file (.memory-sync) inside each mount directory to detect and handle cross-partition conflicts — preventing one team's memory store from silently being misused by another team that happens to share the same directory path.
Feature-flagged (gated by tengu_silk_almanac)
- Each team's memory is scoped to a unique
partitionId(the backend's partition identifier, likely an org or team UUID) - Each mount directory gets a
.memory-syncmanifest (JSON:{ v: 1, partitionId: "..." }) written at sync time - On pull, the manifest's
partitionIdis compared against the expected backend partition; a mismatch invalidates the basis and forces a full re-sync - If a different partition's manifest is detected during a non-user-scoped mid-session sync, sync is suppressed entirely with a warning: "mount dir holds a different partition's
.memory-sync— suppressing sync (remove the dir to re-mount)" - Without this flag enabled, two teams sharing the same directory path could silently read and overwrite each other's memory
New manifest functions (gated on it("tengu_silk_almanac", !1), search for ".memory-sync" or "partition mismatch")
Strings lifted out of the shipped bundle, so the claim above can be checked against them.
tengu_silk_almanac Off in both readingsThe flag server returned off for the account this site reads and for the anonymous baseline. A reading of off cannot rule out a rollout these two readings sit outside of.
This account: off · anonymous baseline: off · compiled default in v2.1.187: off
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 v2.1.187. It isn't a statement about your account. What a flag value here can and cannot tell you
A reading is one sample. Claude Code evaluates its flags remotely, so no client sees the targeting rule behind a value and this says nothing about your account.
A reading is one sample. Claude Code evaluates its flags remotely, so no client sees the targeting rule behind a value and this says nothing about your account.
New in this build: tengu_silk_almanac