Rename init-0.4th..init-9.4th to workload-0.4th..workload-9.4th; fix stale WL-HI/WL-LO doc

These are alternate boot personality capsules, not files doe.4th
dispatches from -- confirmed by reading doe.4th itself (it generates
its own synthetic workload internally, DOE-WORK) and mkcapsule.c
(only the exact filename "init.4th" is special-cased as the active
MAMA_INIT capsule). "workload-N.4th" names them for what they are
without colliding with that reserved name.

Renamed the 10 files (git mv, preserving history) and their own
self-referential header comments (also fixed a pre-existing typo,
"init-4.th" -> "workload-4.4th"), updated capsules/README.md,
capsules/MANIFEST.md (21 references), .claude/CLAUDE.md, and a
tools/mkcapsule.c comment.

Also fixed experiments/bare_metal/README.md's "Adding a Custom
Workload Capsule" section, discovered stale while doing this rename:
it documented a WL-HI/WL-LO dispatch table and a wl_id CSV column
that don't exist anywhere in the current capsules/ tree or doe.4th's
own CSV header -- corrected to describe what's actually there (no
pluggable workload dispatch; a custom workload is run by substituting
it in as the boot's own init.4th).

Verified: mkcapsule --lint clean (35 files, 0 violations), all 3
architectures build with zero new warnings, amd64 boots to zuse)ok>
with an unchanged dict_hash/capsule_hash from every prior boot this
session (0xc8f4b09e36f4fc4a / 0x1ef4939ed32ec1e6) and zero faults.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EXieurDfDSsDFdnSyusuWo

capsules/workload-8.4th

@@ -0,0 +1,16 @@
+Block 2160
+( workload-8.4th - fib )
+: FIB-SEQ
+  0 1
+  100000 0 DO
+    OVER + SWAP
+  LOOP
+  DROP DROP
+;
+: RUN-FIB
+  20 0 DO
+    FIB-SEQ 46 EMIT
+  LOOP
+;
+FIB-SEQ
+( RUN-FIB )