release: v4.0.0-alpha.3; the proofs run capped to fit the server

v4.0.0-alpha.2's release run (200) was killed in the StarForth session:
Isabelle's defaults (Java 4 GiB, a thread per CPU). scripts/nightly.sh
proofs now runs one thread, Java 512 MiB, Poly/ML 900 MiB, through a
settings file of a home made for the run (heaps linked to the user's):
47 s, ~1.8 GiB, logs/20261107-195341. Nothing was published from alpha.2;
its tag stays. LITHOS_VERSION 4.0.0-alpha.3. FABRIC-3.md XLVII.6.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GE3SDvQn5EoxdGufk84zz2

kernel/Makefile

@@ -98,6 +98,10 @@ endif
 #   v4.0.0-alpha.2 — the first published alpha: v4.0.0-alpha.1 (233025da) was accepted and tagged,
 #             but its release run's proofs step built HOL-Library from scratch and was killed, so
 #             nothing was published; the tag stays where it is (Captain Bob, 2026-11-07).
+#   v4.0.0-alpha.3 — the first published alpha: v4.0.0-alpha.2 (2d43db25) kept the image's
+#             HOL-Library, but its proofs ran with Isabelle's defaults (Java 4 GiB, a thread per CPU)
+#             and were killed in the StarForth session; nothing was published, the tag stays. The
+#             proofs now run capped (scripts/nightly.sh: 1 thread, Java 512 MiB, ML 900 MiB).
 # LITHOS_VERSION (kernel) and VERSION (embedded StarForth engine) are independent
 # and do not auto-sync (§XXX.4) — do not infer a relationship if they ever coincide.
 # Engine VERSION bumps at tag time by what actually changed (§XXX.6): major = any
@@ -106,7 +110,7 @@ endif
 # the messaging layer was replaced by kernel-Hermes, PLOT/FB-* relocated to Hestia —
 # all dictionary-visible, no FORTH-79 word semantics changed.
 VERSION       ?= 3.2.0
-LITHOS_VERSION ?= 4.0.0-alpha.2
+LITHOS_VERSION ?= 4.0.0-alpha.3
 
 # ==============================================================================
 # BUILD PATHS