Workspace flow

A workspace is any directory where you ran seif init. It owns a .seif/ store — memory modules, signing material, and the governance ledger — that makes the work inside it durable and verifiable. Everything else in SEIF lives in a workspace.

seif init      →   .seif/ store created
   │               (memory, keys, ledger)
   ▼
seif status    →   identity + counts at a glance
   │
seif doctor    →   read-only health: workspace + runtime install
   │
seif reconstruct → materialize any historical tree on demand

Create

seif init

This writes a SEIF-WORKSPACE-v2 store under .seif/. We refer to the workspace root as $SEIF_CONTAINER_ROOT throughout the docs — it is wherever you chose, never a fixed location.

Inspect

seif status     # identity, module counts, store summary
seif doctor     # one read-only report across workspace + runtime checks

status is the quick glance; doctor is the full diagnostic — it catches runtime-install drift (a stale or shadowed binary) as well as workspace issues, and never mutates anything.

Replay

seif reconstruct <ref>   # materialize a historical workspace tree as a hardlink farm

Because the store is content-addressed, any past state can be rebuilt exactly, without copying the data twice.

Where to go next

• What is SEIF — the mental model behind the store.
• Memory flow — the first thing you put in a workspace.
• CLI reference — every command and flag, always current.