KB: sweep STALE "unfinished" claims -> match landed commits (context-system audit)
4-agent audit cross-checked every open/pending/stub/deferred claim in the context
system against git + code. Fixed ~13 stale spots (currency only, no code change),
each contradicted by its own file body + a landed commit:
- 0xBD3 SubsystemMessageManager (task #7, afefaee): subsystems.md frontmatter+body
line 57; decomp-reference.md 0xBD3 row + +0x434 row -- all still said
"MISLABELS controlsMapper / WAVE 8 untangle" (done; mech.hpp names it messageManager).
- Cylinder hit-location STEP 6 (d07ac7d): combat-damage.md frontmatter;
decomp-reference.md "not reconstructed" line.
- Per-mech turn-rate wiring (task #64b, a83995a): multiplayer.md + locomotion.md
frontmatter (still "bring-up constant / not yet wired").
- Mech-level update records (task #1, c9f0c2a): open-questions.md "dead Wword
branches mech.cpp:1511/1613" (those branches are gone; named fields now).
- Collision damage + DeathShutdown (2026-07-08/12): combat-damage.md frontmatter.
- Myomers un-stub (8b36440): decomp-reference.md "INERT un-stub" row.
- Day/night BT_MATPRI (shipped, default-on): rendering.md frontmatter (LOD half kept).
- Fire-trigger path (task #5, 8ed6184): open-questions.md "still gBTWeaponTrigger
bring-up" (globals retired/never-read).
- HEAT entry FACTUAL fix: "heatLoad -> effectiveRange 0 = overheat cutout" was false
per aab7a8a -- effectiveRange reads host-zone damage; the real cutout is the
FailureHeat charge-hold.
Frontmatter open_questions re-pointed at genuinely-open items. VERIFIED NOT stale
(left as-is): multiplayer body-channel-projection notes (a real open [T3] residual
per the coupled-motion section's ~2.9u leg-vs-body snap). Genuinely-open items
(factory roster loops 2-4, subsystem-panel gate, missile cluster, console-death
freeze, gyro tails, etc.) untouched. checkctx CLEAN.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
