Recover verified USB state when e2fsck preen exits 4

[Deadbytes101]

Alpha.11 physical boot on exact image SHA-256 7870c2c74e9edf64c58372b39fa15db10fe9ad5b605a26d638ea36d87df8d1dd reached the new bounded resize recovery path, but the verified USB state filesystem remained unavailable.

Physical evidence:

• release commit: 15f1a2c0476d00724fc9054347de268d2a9d1162
• state device: /dev/sdb4
• state PARTUUID: 5249474f-04
• state outcome: repair_required
• exact failure: post-timeout ext4 check failed: e2fsck: exit 4: RIGOS_STATE_SEED: UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY.
• /var/lib/rigos remained the live overlay, not the persistent ext4 mount
• recovery credential truthfully remained credential_scope=boot, credential_persisted=false
• SSH truthfully used an ephemeral host identity
• profile and miner remained blocked

Required fix:

• advance the next immutable artifact to Alpha.12
• only after current-boot USB state attestation, exact PARTUUID/ext4/label verification, and proof that the device is unmounted, allow a bounded automatic e2fsck -f -y escalation when preen exits exactly 4
• revalidate the exact state device before escalation and before resize
• retain bounded command output and timeouts
• do not escalate operational, usage, cancelled, or library failures
• rerun the normal state core after successful repair and resize
• keep recovery credentials boot-scoped until state is genuinely mounted and ready
• keep configuration and mining fail-closed

Physical acceptance requires state ready, persistent password restoration across reboot, persistent SSH fingerprint, configuration activation, and miner startup.

[Deadbytes101]

Fix implemented in draft PR #28 at exact head 55dfd228c41177893bf9be4ae6f1d285ac4a11c9.

The state orchestrator now recognizes the core's exact e2fsck exit-4 message, revalidates the current boot-attested unmounted USB state device, runs preen first, and escalates only exit 4 to bounded e2fsck -f -y. The exact device is revalidated again after repair and before resize. Other e2fsck failure classes remain fail-closed.

Alpha.12 source verification, exact image construction, and physical persistence tests remain pending.

[Deadbytes101]

Alpha.12 source regression gate passed at exact head 5af491daed09dff4b5f90244f4fee718edd5ade2.

The dedicated state recovery tests passed for the bounded verified path, including preen-first behavior, escalation only for e2fsck exit 4, no escalation for operational failures, revalidation before resize, and truthful failure reporting. The full WSL source gate ended with RIGOS_WSL_SOURCE_GATE=PASS.

Exact image construction and physical recovery of /dev/sdb4 remain pending.

[Deadbytes101]

Alpha.12 exact image verification is now PASS at exact head 5af491daed09dff4b5f90244f4fee718edd5ade2.

Artifact:

• dist/usb/rigos-usb-amd64-0.0.4-alpha.12.img
• SHA-256: e4cc44215c70cba7b3d7a9c5916256992c40161e110314a657c1856d9d1a3a16

All exact-image gates passed, including the new state repair and credential verifier. Physical validation is still required for the original exit-4 failure: /dev/sdb4 must repair, resize, mount at /var/lib/rigos, persist the recovery credential and SSH identity across reboot, then allow configuration and miner startup.