BFM-decomp/phase-ends/CURRENT_PHASE.md

CURRENT_PHASE — Phase 32: the last 21 + the parked 5 (short, kill-gated)

Gate 1 approved 2026-09-05 (Drew, plan mode, Max, Fable 5.1). Rules R44–R63 ratified at gate 1 (binding). The approved plan (full text): ~/.claude/plans/plan-mode-enabled-max-iterative-ember.md — its substance is this file. Jumping-off document: docs/frontier-p32.md. Roadmap: docs/roadmap-to-100.md v2 §3 (P32 as written is consumed; P32 is this census) + PhaseEnd_Phase31.md Roadmap delta. Baseline HEAD at open: commit:3900 (Phase-31 close + post-close docs; no src//config/ change since the S80 census — verified git diff commit:3898..HEAD -- src config Makefile empty).

Milestone (gate 2)

Every one of the 21 census rows is BANKED or carries a FINAL verdict (PROVED wall / candidate wall with a gcc-mechanism citation + a bounded-attempt record / deferred with its cost stated); each of the 5 parked payloads is onboarded byte-identical at a byte-proven base (its stubs then join the census and are banked or ledgered the same way) or carries a documented exclusion with the static + runtime evidence; the fleet is byte-identical (R22) at every banked batch; the PhaseEnd carries the final wall ledger, the corrected denominators (main game-code 41,556 — the 22-ins Ghidra-boundary gap), all three metrics, and the roadmap delta (P33 = verify + flip).

Kill gate: a session that banks nothing and produces no new verdict ends the crack work (T3); the parked-5 work is bounded to the static candidate list + one static-RE session; no forced banks, no redefinition (P9). Scale estimate: 3–5 sessions.

Effort / model (R7/R26/R27 — every transition is PROMPTED, never assumed)

• Max through T0–T2 (T2 is non-obvious static RE; T0/T1 are settled-design and would be fine at xHigh).
• Prompt to drop to xHigh for T3 (orchestration) and T4 (ledgering); prompt back to Max for T5 (PhaseEnd).
• Models: T1c one Sonnet agent (123 ins); T3 Opus for all 11 targets (escalation tier — Sonnet plateaued on the ≤120-ins rows in S79); Fable only if an agent names a wall class the codegen map lacks — STOP and prompt Drew first. No Ultracode / Workflow: one-page census → the S80 one-agent-per-function shape (docs/wave-playbook.md §S80 addendum, Agent-tool subagents; tools/agent_verdicts.py for any that outlive a session).

Adjustments vs frontier-p32.md §4 (approved)

1. The parked 5 (T2) run BEFORE the crack pass (T3): the static header probe (2026-09-05) gave every payload a deterministic candidate base, so T2 is a set of ~0-token new_binary.sh byte-gate probes, and any stubs they add ride the same single bounded agent pass.
2. Tooling-first split into three tasks, one commit each (P4/R42/R60).
3. No MCP dependency for the parked 5 (static evidence + our own matched C); MCP only as a fallback (R29 prompt).
4. Walls get a ledger task (T4), not drafting — incl. fixing two dangling best-draft paths (R62).

Tasks (strict order; one commit per task after this file is updated; count banks from the SOURCE)

• T0 — Phase open (Max) — DONE 2026-09-05: verbatim 5==5 · exclude 8/8 fresh · census 21 (B-CARVE 4 / D-NEAR 10 / F-FAR 7, .run/P32/frontier_p32_open.json) · tools-health OK · check-all 213/213 rc 0 (.run/P32/t0_baseline.log): this file; harness task list (R28); baseline reads — verbatim_check --strict (5==5), exclude_audit --assert-fresh, frontier_classify --json .run/P32/frontier_p32_open.json (21), make tools-health OK, make check-all 213/213 (read of the on-disk S80 artifacts); commit.
• T1a — resident: split the code subseg, bank func_800D128C (243) — DONE 2026-09-05: BANKED byte-identical 8e17e02f… (commits commit:3902 split + commit:3904 bank); R22 after the bank: check-all 213 passed / 0 failed, rc 0 (.run/P32/t1a/r22.log); resident stubs 2 → 1; THREE instrument fixes on the way (cookbook §496 carrier provided-types, §498 carve --pre drop + gate ignoring the extract rc + interleave_check false DRIFT); twin_rescan: 20 open, nothing newly free. Original brief: jr_isolate_all resident --only func_800D128C (dry-run CLEAN 2026-09-05: 3 region files, carve repoint resident → resident_jr_800D00E4, --order + JTBL_PADS) → make extract BINARY=resident && make build BINARY=resident -j8 (EXIT CODE, R53; byte-neutral split) → commit carve state alone (R60) → rtu_match the closeness-0 draft .run/S71_gate14/resident/func_800D128C.c in the new region TU → parallel_gate --r22 (carves jtbl_80113FB8 + jtbl_80114198; pads_audit after) → in-tree build exit code → commit → twin_rescan. TU refusal → §376/§378 chain as a byte-neutral plumbing commit BEFORE re-gating.
• T1b — ov_SC02_017: split + remap the twin, bank func_80186C64 (209) — DONE 2026-09-05: BANKED byte-identical c0253499… (commit:3903 carrier §497 fix — NO source rename needed · commit:3906 split · commit:3907 bank); 0 drafting tokens; four §376 TU spellings + one TU-provided typedef stripped. Original brief (Max/xHigh): byte-neutral tag rename in src/ov_SC02_017/ov_SC02_017_jr_8017DF34.c (:4314/:4319; keep the typedef name) → build sha-unchanged → commit → jr_isolate_all ov_SC02_017 --only func_80186C64 (dry-run REFUSED on the conflict 2026-09-05; expect CLEAN after the rename) → extract/build byte-neutral → commit carve state → family_remap --addr 0x801810C8 --from ov_SC02_016 --to ov_SC02_017 --to-addr 0x80186C64 (d=2 RELOC-ONLY twin; source src/ov_SC02_016/ov_SC02_016_jr_801810C8.c:2839) → strip typedefs the TU already defines (§491 gap 2) → rtu_match → parallel_gate --r22 → commit → twin_rescan.
• T1c — md_MAIN_034: bank func_800CB00C (123) — DONE 2026-09-05: BANKED byte-identical 46153c06… (commit:3908) from the STORED S72 body .run/O21/opus/func_800CB00C.c (rtu MATCH 123/123; the census's best_draft was a name-collided WRONG file, R48/R38) — NO agent; WALL pin dropped (exclude 8→7); backlog 18→17. R22 after the T1b+T1c batch: check-all 213 passed / 0 failed, rc 0 (.run/P32/t1b/r22.log). Original brief: pack via claude_wave_packs (the stored 174-ins draft is WRONG, not a seed; §494 jtbl idioms) → verbatim grep → rtu_match --split md_MAIN_034 --source md_MAIN_034 → parallel_gate (island pads derive at build, §303) → commit → twin_rescan; drop the md_MAIN_034:func_800CB00C pin (exclude_audit --write). Progress report to Drew after T1c.
• T2 — the parked 5 (Max):
  • T2a — DONE 2026-09-05 (commit:3909): controls 7/7 (5 STRONG, 2 CONSISTENT; TEXT_LO estimates == yaml); the five: MAIN/7 STRONG 0x800CEDF8 · MAIN/9 STRONG 0x800CD348 · SC03/53, /54 CONSISTENT top 0x801EF468 (12-way tie) · SC03/56 SPLIT (jal vote 0x80178C8C vs ptrs/lui ~0x801CBB50); memory-map §S45 p7; SETUP row. Brief: tools/payload_base_evidence.py (id word, first prologue → TEXT_LO, abs-pointer window, lui hi-half histogram, jal self-consistency per candidate base, DESTPTR lookup in the window from .run/idxtab_map.json) — R39 controls: must re-derive md_MAIN_008/011 (0x800CEDF8), md_MAIN_042 (0x800CCB1C), md_MAIN_013 (0x800CAE08), md_SC03_073 (0x801EF468) from their payloads alone; refuses a payload with no self-reference (base-independent bytes — say so, never guess). SETUP row (R21); evidence rows → docs/memory-map.md §S45 p7 (G5 static-derived, US). Commit.
  • T2b — DONE 2026-09-05: ALL FIVE onboarded on their first candidate, byte-identical first builds — md_MAIN_007 @0x800CEDF8 (commit:3911, 19 stubs, A4 applied) · md_MAIN_009 @0x800CD348 (commit:3912, 11) · md_SC03_053 + md_SC03_054 @0x801EF468 (commit:3913, 15 + 7) · md_SC03_056 @0x801CBB50 (commit:3914, 4). Fleet 213 → 218; +56 stubs (~2,600 ins). Instrument findings (R34/R35): the all-INCLUDE_ASM first build is a NULL oracle for FINE base errors (+8 builds byte-identical; +0x1000 fails the link) — bases are static-derived STRONG, byte-proven by the first internal-call C bank; a jal-vote base whose targets are shared-engine function starts is OUTWARD-EXPLAINED (SC03/56's false 0x80178C8C). Close chain (sig-modules → tools-health → audit-disc → R22 of 218) running: .run/P32/t2b/close.log. Original brief: onboard probes, one payload per commit, top candidate first (the module-class law: the FIRST build's byte-identity is the arbiter; a miss → next candidate; none → stays parked with the evidence): md_MAIN_007 extracted/retail/MAIN.CD.dir/FILE_007 0x800CEDF8 0x34 (9/9 internal jals on its own prologues at that base; omit symbols.resident.txt, A4 law) · md_MAIN_009 …/FILE_009.dir/0.1 0x800CD348 0x3C (the unique base fitting both in-region jals; then 0x800CAE08, 0x800CCB1C) · md_SC03_053 …/SC03.CD.dir/FILE_053.dir/0.1 0x801EF468 0x4 · md_SC03_054 …/FILE_054.dir/0.1 0x801EF468 0xF0 · md_SC03_056 …/FILE_056.dir/0.1 0x801CBB50 0x4 (then 0x801CBD90, 0x801CC6E0). After each success: make sig-modules, make tools-health (R36), make audit-disc (UNCLAIMED 5 → 5−N), commit.
  • T2c — DONE 2026-09-05: 20 free banks in the five new modules for 0 drafting tokens (commit:3918 8 · commit:3919 5 · commit:3920 7): 7 d=0 hash twins + 3 RELOC-ONLY via family_remap, 9 constant-only NEAR-COUSINs (each constant named by the real-TU diff), 2 hand pass-throughs (main's func_80014128 shape); dedup_extend 0 (no macro-group members); split_indicator population fix (commit:3917, was a stale 213 list — R51). R22 after the batch: check-all 218 passed / 0 failed, rc 0 (.run/P32/t2c/r22.log). Census: 54 stubs / 6,376 ins (.run/P32/frontier_t2c_close.json: main 12 · md_MAIN_007 15 · md_MAIN_009 10 · md_SC03_053 6 · md_SC03_054 4 · md_SC03_056 1 · ov_SC03_105 2 · resident 1 · md_MAIN_003 1 · ov_SC06_022 1 · ov_SC07_002 1). make audit-disc UNCLAIMED 0 of 220; disc-completeness.md P32 section written; cookbook §499 + decision-log S81 (R31). Brief: free banks + census: twin_rescan / dedup_extend for h_exact twins of the new stubs; regenerate the census; leftovers join T3. Update docs/disc-completeness.md (+ ledger regen), decision log (R31).
  • T2d — NOT NEEDED (both MAIN/7 and MAIN/9 onboarded on their first candidate). Brief: (only if MAIN/7 or MAIN/9 refuse every candidate): one bounded static-RE pass over our own C (src/800.c, src/boot.c, src/resident/resident.c: loadDestPtrTable, D_800D3764/D_800D384C, the resourceIdMap sites) for a route producing global index 7 or 9; MCP (R29) only if the C is insufficient. Outcome: a base, or the strengthened exclusion evidence. Progress report after T2.
• T3 — one bounded crack pass, Opus agents, one per function (xHigh orchestration — prompt R27): IN PROGRESS — 47/47 drafted, 38 BANKED (10 in the launching session + 28 in S83 = 2,111 + ~2,800 ins; commits commit:3923..commit:3931, commit:3935..commit:3946), 9 NEAR ledgered in the backlog, 0 FAIL outstanding; the 17 queued Haiku rows went 17/17 MATCH (S83). Census 16 stubs / 3,827 ins = 7 pinned walls + 9 NEAR (.run/P32/frontier_t3s3_haiku.json); md_SC03_053/054/056 100% C; resident 145/145. Remaining in T3: step 8 (bounded NEAR tail: permuter_ils ×2 + one Opus look) and the close (step 9). Earlier checkpoints: 09:30 / 10:01 MDT (the launching coordinator overflowed at 07:07Z; the S82 successor recovered 22 verdicts). Original brief: targets = main:func_80023BF0 11 · func_80038698 11 · func_80039B20 7 · func_80039308 34 · func_80015B6C 40 · func_8002FDE8 35 · func_8001BC6C 28 · ov_SC03_105:func_80185810 37 · md_MAIN_003:func_800CF3E8 54 · ov_SC07_002:func_8017DC80 84 · resident:func_800D06E8 344 (after a byte-neutral fix of the Struct80078E78 typedef clash) + any T2c leftovers. Cheap second looks first: func_80023BF0 (VERIFY the S80 permuter waypoint's semantics before seeding) and func_80039B20. Packs: claude_wave_packs (S79 verdict incl. INERT list, exact-length best draft, matched same-TU neighbour, §494/§495 laws, the fixed permuter recipe permuter_ils … --cycles 8 --secs 150 --j 3); brief = .run/S79w/SYS.md. Per result, never in bulk: verbatim grep → rtu_match (main: --tu src/<TU>.c) → gate (parallel_gate --r22 / main: gate_main <slate> --apply, read the BODY/TABLE/PLUMBING/MIXED line) → commit per bank (R42) → twin_rescan; plateaus → backlog.py log with attempt + cost (R41). Expected 1–4 banks (S79: 3/11 on the Opus F/G tier), ~0.6–1.4M tokens. Progress report after T3.
• T4 — the walls' FINAL verdicts (xHigh; no drafting): the 8 pinned rows — confirm each citation current + instrument exonerated (R40), PROVED vs CANDIDATE, ledger best-draft → an existing real-TU-clean C file (R62: func_80011380 → .run/m3/opus/func_80011380.c; func_801834A4 → .run/S71_gate14/ov_SC03_105*/), exclude_audit --write, backlog.py render; a T3 idiom naming a wall's mechanism re-opens that row (bounded: one attempt, permuter first). Wall table into this file; cookbook §496+; decision log; accelerators.
• T5 — PhaseEnd (Max, Tier 1 — prompt R27): P7 checkbox walk; milestone demo (R22 fleet N/N, tools-health, verbatim_check --strict, final census, wall ledger, parked-5 dispositions via make audit-disc, make report all three metrics + main 143dbb89… with/without SDK dirs, corrected denominators); WAIT for gate 2; PhaseEnd_Phase32.md (+ Roadmap delta + Plain-English Recap); git mv this file → phase-ends/logs/Phase32.md (R19); R23 MCP stop if it ran; leave both uncommitted for Drew (R6); 🛑 (P8).

Standing procedure (every task)

Fix the TU by a byte-neutral plumbing commit BEFORE any gate (S77 law) · commit banked work before the next command that can touch src/ (R42) · after any bank: twin_rescan, verbatim_check --strict (R62) · after a carve/merge touching a yaml/.mk: make extract BINARY=<b> before building; pads_audit/interleave_check on every touched carve binary (R60) · fleet verify after every banked batch: make clean && make extract-all && make check-all — read EXIT CODES (R53), quote the green count at every close (R58); make tools-health · log each task here + refresh the 🛑 checkpoint before any pause · rules check every 4 tasks (P6) · harvest idioms into the cookbook before the next drafting step (R16/R30) · no Co-Authored-By trailer (R5); keep the Claude-Session: line.

Blockers

• 2026-09-05 09:xx: the successor session's SessionStart hook launched a headless Ghidra MCP server (.run/ghidra-mcp.log); T3/T4 do not use it; no RE writes → never stage the ghidra/ churn (R23).
• Ghidra MCP not running at open (clean stop in the log; the SessionStart hook's server did not persist). Not needed for T0–T2b; if T2d needs it: tools/ghidra_mcp_start.sh → pause → Drew runs /mcp (R29) → G2 ping.

Log

• 2026-09-05 10:10–11:25 MDT (S83, session 491895ad, xHigh, Fable 5.1) — T3 steps 0–7 DONE: 28 banks, fleet R22 218/218 twice. Preflight R22 218/218 rc 0 (.run/P32/t3s3/r22_check.log) → the 11 S82 MATCH rows banked: md_SC03_053 ×2 (commit:3935 commit:3936, 100% C) · main ×2 via gate_main (commit:3937: "slate 2 -> 2 compatible … BANKED 2 … 143dbb89 BYTE-IDENTICAL") · md_SC03_054 func_801EF6D8 604+7 jtbls (commit:3938; jtbl_carve --probe refused the tail carve → §303 derive stage reproduced the island pads 0,0t1,0t1,0t1,0t1,0t1,0 with NO carve state change) · md_MAIN_007 ×6 after the byte-neutral func_800CF3B0 no-proto decl commit (commit:3939, commit:3940) — two integration classes on the way (same-TU s16/u16 spelling of D_800B99E8; §304 self-defining rodata D_800CEDFC). twin_rescan 33 open / 0 free; verbatim 5==5; R22 make clean && make extract-all && make check-all → 217+main extracted, 218 passed / 0 failed, exits 0/0/0 (.run/P32/t3s3/r22_full.log, 10:41–10:46); tools-health OK. The 9 NEAR rows ledgered (backlog_near.sh; docs/backlog.md 15 open). 17 Haiku agents launched 10:57 from staged prompts (.run/P32/t3s3/prompts/), 17/17 MATCH by 11:15, each re-verified with rtu_match in the CURRENT TU and banked by bank.sh (verbatim grep → rtu → splice → one build → sha → commit-on-green): md_MAIN_007 ×6 (commit:3941 commit:3942), md_MAIN_009 ×8 (commit:3943 commit:3946 — func_800CD520 needed a §376 re-spell after a sibling bank added the file-scope prototype), md_SC03_054 ×2 (commit:3944, 100% C), md_SC03_056 ×1 (commit:3945, 100% C). twin_rescan 16 open / 0 free. Harvest: cookbook §500-F (the four integration classes), BRIEF.md §304 sentence, jtbl_carve --probe text, harvest_notes. Post-batch R22 (after all 28 banks): 217+main extracted, 218 passed / 0 failed, exits 0/0/0 (.run/P32/t3s3/r22b_full.log, 11:04–11:09).
• 2026-09-05 (successor session, later) — Session-start protocol rewritten at Drew's direction (R64 candidate): phase-ends/DIGEST.md created (every phase synopsis + every rule in full + the PROJECT_CONTEXT corrections + the doc map); CLAUDE.md load order = PROJECT_CONTEXT → DIGEST → the three most recent PhaseEnds → CURRENT_PHASE, ≈100k tokens, the 🛑 block replayed VERBATIM; Phase Boundary step 3b maintains the digest (P7); phase-ends/README.md + SETUP §7 updated; the two memories (checkpoint-current-phase-before-pause, session-start-list-rules-in-full) + MEMORY.md updated; .run/P32/t3/PROMPT_TEMPLATE.md saved (verbatim agent prompts). The 🛑 block below was REFRESHED to the verbatim-replay standard and SUPERSEDES the 09:30 block. Ratify R64 at PhaseEnd_Phase32.
• 2026-09-05 09:30 MDT — T3 CHECKPOINT by the successor session (Max, Fable 5.1) after the launching session died. The T3 coordinator (session c6a5fb71) banked 10 functions in 9 commits (T3 (1)–(9), 01:01–01:07 MDT) and then hit "Prompt is too long" at 01:07 MDT; 22 agent completions (07:07Z–08:04Z) arrived into a dead session. Recovered via the recorded route: agent_verdicts.py over the 31 task transcripts → .run/P32/t3/verdicts.jsonl (20 MATCH / 9 NEAR / 2 FAIL); every unbanked draft re-verified with rtu_match in its real TU (10 MATCH confirmed). Instrument incident (R48 class): one agent's tidy-up (find .run/P32/t3/opus -maxdepth 1 -type f ! -name func_800CD674.c -exec mv {} _scratch/, 07:52Z) had swept 11 sibling Opus deliverables — found in _scratch/, moved back, byte-verified at the reported closeness; tools/agent_drafts_restore.py (NEW — transcript replay) written as the fallback; .gitignore allowlist for .run/P32/** so the drafts/ledger/census are committed (R20). Harvest written IN-SESSION (R16/R30): cookbook §500 (10 banked closers, 10 MATCH closers, 9 NEAR classes, two NEW mechanisms — the pinned-base alias basin and #line-equalised ASM_OPERANDS — and the process defects); docs/wave-playbook.md §S80 addendum-2 (per-function work dirs, JSON-only final message, the 20-agent cap, the recovery tools); docs/accelerators.md P32 T3; docs/decision-log.md P32 S82 (R31); SETUP row (R21). Read-only T4 pre-check results from the dead session preserved in the 🛑 block. Effort: this checkpoint at Max; T3 resumes at xHigh (R27).
• 2026-09-05 — gate 1 approved; T0 opened at HEAD commit:3900.
• 2026-09-05 — T3 LAUNCHED (xHigh): 47 targets = census 54 − 7 pinned walls (.run/P32/t3/targets.json, 5,288 ins); packs built (claude_wave_packs, 47/47 neighbours, 10 journals); brief .run/P32/t3/BRIEF.md. 20 Agent-tool drafters running (16 Opus: the 11 old near/far rows + resident func_800D06E8 + the 4 largest new bodies · 3 Sonnet · 1 Haiku); 27 Haiku queued behind the harness's 20-concurrent cap (.run/P32/t3/pending_launch.txt, launched one per freed slot). Recovery + per-result route: .run/P32/t3/LAUNCHED.md.
• 2026-09-05 — T2c DONE (R22 218/218 rc 0) · T2d not needed. Twenty of the 56 new stubs fell to twin remaps and constant flips (0 tokens); the rest (36, all ≤50 ins but one, never drafted) join T3 on the model ladder. Session limit reset here (Drew); effort now xHigh via /model (the T3 orchestration level, R27 satisfied). Settings-file JSON defect fixed (a trailing comma).
• 2026-09-05 — T2b DONE — the parked ledger is EMPTY: all five payloads onboarded byte-identical (MAIN/7 → md_MAIN_007 @0x800CEDF8; MAIN/9 → md_MAIN_009 @0x800CD348; SC03/53, /54 → md_SC03_053/054 @0x801EF468; SC03/56 → md_SC03_056 @0x801CBB50). Every base came from T2a's static alignment (self-calls / fn-ptr-table entries on the module's own function starts; controls 7/7), NOT from the byte check — two negative controls showed the first build is base-lenient (fine offsets) and only catches gross errors. Two tool refinements en route (combined STRONG rule; outward-explained vote bases). The 100% claim loses its last asterisk once the new stubs are banked or ledgered (T2c/T3).
• 2026-09-05 — T1b + T1c DONE (R22 213/213 rc 0 for the batch) · T2a DONE. Both banks were ZERO-token: T1b's body came from family_remap of the d=2 twin (+ four TU spellings), T1c's from a stored S72 body the census had mis-keyed (R48) — the journals (R38) named both. Census 21 → 18 stubs (resident 1 · main 12 · md_MAIN_003 1 · ov_SC03_105 2 · ov_SC06_022 1 · ov_SC07_002 1). T2a's instrument gives every parked payload a bounded probe list; two are STRONG.
• 2026-09-05 — T1a DONE — resident:func_800D128C BANKED (243 ins, byte-identical 8e17e02f, R22 213/213). The stored S71 closeness-0 draft was byte-correct all along; the whole task was three instrument defects the resident (the fleet's one common.h-only, --pre-sandwich binary) exposed in overlay-only assumptions: (1) jr_isolate_all dropped a file-local typedef whose name engine_types.h also defines (§496 — fixed: provided types derived from the TU's own includes); (2) jtbl_carve regenerated JTBL_INTERLEAVE without the --pre hdr.rodata.o clause → extract refused → the gate linked a stale script and booked the byte-correct draft as DIFF; harvest_verify ignored that extract's rc (§498 — both fixed, R49/R61); (3) interleave_check read a --pre line as n=0 (false DRIFT; fixed). R38 then found two more stored MATCH bodies for T1b/T1c (see their rows). Effort stayed Max.
• 2026-09-05 — T0 DONE. Baseline reads all green (.run/P32/t0_baseline.log): verbatim_check --strict 5 bodies == 5 rows; exclude_audit --assert-fresh 8 entries, 8 WALL, 0 stale; frontier_classify → 21 rows = the S80 census exactly; make tools-health OK (sigs fresh, corpus(+resident), cdecl, audit-binaries 213/213, report lint+dedup, cookbook-index, split_indicator 213 OK); make check-all 213 passed / 0 failed, rc 0. Harness task list #1–#11 built (R28). NEXT = T1a.

🛑 SESSION CHECKPOINT — T3 STEPS 0–7 DONE, STEP 8 NEXT (2026-09-05 11:30 MDT; written by session 491895ad "S83"; SUPERSEDES the 10:01 block)

0. How to use this block

You are a FRESH SESSION that has read PROJECT_CONTEXT.md, phase-ends/DIGEST.md, PhaseEnd_Phase29/30/31.md and this file, and nothing else (CLAUDE.md protocol, R64 candidate). Replay this block verbatim into your chat, state phase / done / NEXT / effort, list the rules from the digest, then WAIT for Drew. When he confirms, follow §4 in order. Effort: xHigh for the rest of T3 and for T4 (R27 — Drew sets it; if the status line does not say xHigh, prompt before the first command); Max only for T5. Fable only if a wall class the codegen map lacks appears — STOP and prompt Drew first. No Workflow / Ultracode: one Agent-tool subagent per function where an agent is used at all.

1. Where we are

Phase 32 — the last 21 + the parked 5 (short, kill-gated). Gate 1 approved 2026-09-05 (Drew, plan mode, Max, Fable 5.1); R44–R63 ratified then; R64 is a candidate (DIGEST §3). Tasks: T0 ✓ commit:3901 · T1a ✓ commit:3902 commit:3904 · T1b ✓ commit:3903 commit:3906 commit:3907 · T1c ✓ commit:3908 · T2a ✓ commit:3909 · T2b ✓ commit:3911 commit:3912 commit:3913 commit:3914 (fleet 213 → 218) · T2c ✓ commit:3917 commit:3918 commit:3919 commit:3920, close commit:3921 · T2d not needed · T3 steps 0–7 ✓ (launch commit:3922; banks 1–9 commit:3923 commit:3924 commit:3925 commit:3926 commit:3927 commit:3928 commit:3929 commit:3930 commit:3931; S82 checkpoints commit:3932 commit:3933 commit:3934; S83 banks 10–37: commit:3935 commit:3936 commit:3937 commit:3938 commit:3939(plumbing) commit:3940 commit:3941 commit:3942 commit:3943 commit:3944 commit:3945 commit:3946; the S83 checkpoint commit = HEAD) · T3 step 8 (bounded NEAR tail) and step 9 (close) pending · T4, T5 pending. Harness task list #1–#12 exists (#6 §5.5 and #8 §5.7 in_progress at this writing; #9 = step 8 next). Fleet 218 binaries (main + resident + 138 ov_* + 78 md_*). Last full R22: 218 passed / 0 failed, exits 0/0/0 at 10:46 MDT (.run/P32/t3s3/r22_full.log, BEFORE the 17 Haiku banks); post-batch R22 (after all 28 S83 banks): 217+main extracted, 218 passed / 0 failed, exits 0/0/0 at 11:09 MDT (.run/P32/t3s3/r22b_full.log). tools-health OK at 10:57 (.run/P32/t3s3/tools_health.log). Non-ghidra tree clean at HEAD; the ghidra/ churn (db.*.gbf, ~index.*, *.prp) is R23 restart-noise — never stage it. Claude commits, Drew pushes (R6; ~45 unpushed at this writing).

2. What S83 did (session 491895ad, 2026-09-05 10:10 → 11:30 MDT, xHigh, Fable 5.1)

Followed the 10:01 block's §5 in order. (0) Preflight: tree clean, verbatim 5==5, make check-all 218/218 rc 0. (1)–(4) The 11 S82 MATCH rows banked — each re-verified with rtu_match in the real TU first: md_SC03_053 func_801EF734 44 + func_801EF7E4 72 (→ 100% C); main func_80015B6C 120 + func_8002FDE8 73 via gate_main ("slate 2 -> 2 compatible, 0 dropped … BANKED 2 main functions -- 143dbb89 BYTE-IDENTICAL", .run/P32/t3s3/gate/gate_main2.log); md_SC03_054 func_801EF6D8 604 code ins + 7 jump tables — jtbl_carve --probe REFUSED the tail carve (tables in the §154-A leading island 0x4..0xF0) and named §260, but for an md_* module §303 applies: splice + build, the Makefile's jtbl_rodata_pads --derive stage printed 0,0t1,0t1,0t1,0t1,0t1,0 and the sha held with NO carve-state change (probe text amended); md_MAIN_007: the byte-neutral no-proto decl commit for func_800CF3B0 (commit:3939, sha unchanged) then six drafts — the batch exposed two integration classes rtu_match cannot see (it splices ONE draft): same-TU s16/u16 spellings of D_800B99E8 (fixed to the LOAD's u16), and §304 self-defining rodata (D_800CEDFC = .asciz "C:\TIMPACK\OPDEMO0.PAT", defined const char[] at file scope at the extern's position, (void *) cast at the call). (5) twin_rescan 33 open / 0 free; verbatim 5==5; R22 make clean && make extract-all && make check-all → 217+main extracted, 218 passed / 0 failed, exits 0/0/0; make tools-health OK. (6) The 9 NEAR rows ledgered via .run/P32/t3s3/backlog_near.sh (docs/backlog.md 15 open). (7) The 17 queued Haiku rows dequeued, prompts staged (.run/P32/t3s3/prompts/<fn>.txt = PROMPT_TEMPLATE + two integration sentences) and launched at 10:57 as 17 Agent-tool subagents (model haiku); 17/17 MATCH by 11:15 (43–57k tokens, 46–126 s each); every verdict re-run by the coordinator with rtu_match in the CURRENT TU and banked with .run/P32/t3s3/bank.sh (verbatim grep → rtu ×N → splice ×N → ONE build → sha vs config/check → commit only on green): md_MAIN_007 ×6 → 2 open (both NEAR), md_MAIN_009 ×8 → 2 open (both NEAR; func_800CD520 needed a §376 re-spell of func_8001AD38 to the TU's (const char*) after a sibling bank added the file-scope prototype), md_SC03_054 ×2 (→ 100% C), md_SC03_056 ×1 (→ 100% C). twin_rescan 16 open / 0 free. Harvest (R16/R30): cookbook §500-F (yield + the four integration classes), .run/P32/t3/BRIEF.md (the §304 sentence + exact-TU-spelling), .run/P32/t3/harvest_notes.md, tools/jtbl_carve.py probe text (§303 for md_*), .run/P32/t3/verdicts.jsonl (+17 rows, "session": "491895ad"), .gitignore allowlist for .run/P32/t3s3/. Nothing under config/ changed this session; src/ changed only by the banks listed above.

3. Census — 16 stubs / 3,827 ins (.run/P32/frontier_t3s3_haiku.json; 7 pinned walls + 9 NEAR, 0 undrafted)

binary	stubs · ins	rows
main	7 · 1,450	NEAR (backlog): func_80039308 518 → 17 (.run/P32/t3/opus/func_80039308.c, TU src/800_c.c) · func_8001BC6C 69 → 6 REGALLOC-PERM $v0↔$v1 (…/opus/func_8001BC6C.c, TU src/800.c) · WALLS (pinned, config/wave_exclude.txt, T4): func_80011380 192 (boot −O0; .run/m3/opus/func_80011380.c DIFF 6) · func_80032A74 422 (800_b_2; census closeness 1 from a stored draft — the S79w draft is a CC1 FAIL in the TU, §376 re-probe first) · func_80020DA4 100 (.run/S79w/sonnet/func_80020DA4.c DIFF 2) · func_80039DEC 74 (S79w drafts CC1 FAIL — def-side decl, §495) · func_800391D4 75 (S79w sonnet draft CC1 FAIL conflicting types for D_80073140)
md_MAIN_009	2 · 421	NEAR func_800CD674 174 → 2 REGALLOC-PERM $a3↔$t1 (…/opus/func_800CD674.c) · func_800CD92C 247 → 15 §S7 prologue weave (…/opus/func_800CD92C.c)
md_MAIN_007	2 · 427	NEAR func_800CF408 178 → 49 [permuter] 3 hunks · func_800CF6D0 249 → 137 sched1 class tie (…/opus/)
md_MAIN_003	1 · 469	NEAR func_800CF3E8 469 → 27 pinned-base alias basin §500-D1 (…/opus/func_800CF3E8.c) — the ONE untested lever: an unpinned alias of p6 for the tag load alone
ov_SC03_105	2 · 595	NEAR func_80185810 489 → 35 [permuter] 4 emission windows (…/opus/func_80185810.c) · WALL (pinned) func_801834A4 106 (.run/S71_gate14/ov_SC03_105*/func_801834A4.c DIFF 6)
ov_SC06_022	1 · 119	WALL (pinned) func_8017DF28 119 (.run/S79w/sonnet/func_8017DF28.c DIFF 2)
ov_SC07_002	1 · 346	NEAR func_8017DC80 346 → 46 reload-slot frame + la $a0 slot (…/opus/func_8017DC80.c)

Modules at 100% C after S83: resident (145/145), md_SC03_053, md_SC03_054, md_SC03_056. The 38 T3 banks: the 10 of the launching session (§500-A) + S83's 28 (git log --oneline commit:3934..HEAD). Backlog: docs/backlog.md 15 open near-misses (the 9 NEAR rows carry class · closeness · best draft · cost "1 Opus agent, 30–70 min", R41).

4. NEXT — in order (R42: commit before the next command that can touch src/; a bank that lands goes through bank.sh)

0. Preflight: git status --short | grep -v ghidra/ (empty) · .venv/bin/python tools/verbatim_check.py --strict (5 == 5) · make check-all → 218 passed, 0 failed of 218, exit 0 (the R22 baseline before any bank; if any binary is red, fix the baseline first — R56).
1. T3 step 8 — the bounded NEAR tail (xHigh; the kill gate stands: no bank + no new verdict ends T3): (a) permuter_ils on the two REGALLOC-PERM seeds, §494 recipe, run in the background with logs (R55): .venv/bin/python tools/permuter_ils.py func_800CD674 --draft .run/P32/t3/opus/func_800CD674.c --asm-subdir asm/md_MAIN_009/nonmatchings/md_MAIN_009 --klass REGALLOC --cycles 8 --secs 150 --j 3 > .run/P32/t3s3/ils_func_800CD674.log 2>&1 and the same for func_8001BC6C (--draft .run/P32/t3/opus/func_8001BC6C.c --asm-subdir asm/nonmatchings/800 — confirm the asm dir from corpus.stubs('main')/targets.json asm field before running). Read every reported score's DIFF before calling it a closeness (R63); a 0 → rtu_match in the real TU → bank.sh (main: gate_main slate instead). (b) ONE Opus second look at func_800CF3E8 (md_MAIN_003, 469 → 27) with the single untested lever named in §500-D1 / the backlog row (an unpinned alias of p6 for the tag load alone); prompt shape = PROMPT_TEMPLATE's Opus row + .run/P32/t3/reports/func_800CF3E8__opus__*.md as the prior; draft path .run/P32/t3/opus/func_800CF3E8_s83.c (never overwrite the ledger's best draft); the agent's final message is the JSON line only. Plateaus stay in the backlog (tools/backlog.py log … --closeness N), each with its cost (R41).
2. T3 step 9 — close: any new idiom → cookbook §500-G (R16/R30) + harvest_notes.md; make report (regenerates docs/progress*.md, docs/difficulty.md); frontier_classify --json .run/P32/frontier_t3_close.json; refresh this block; commit; progress report to Drew; mark harness #8/#9/#10 done.
3. T4 (xHigh, no drafting) — §6 below. 4. T5 (Max, Tier 1 — prompt R27; WAIT for gate 2) — §6 below.

5. Files, tools, exact invocations, gotchas

• .run/P32/t3s3/ (S83; tracked: *.sh *.py *.log *.txt prompts/ gate/*.json gate/*.log; untracked: verify*/ compile dirs): splice.py <tu> <asmdir> <fn> <draft> (asserts the INCLUDE_ASM line occurs once, replaces it with the draft) · bank.sh <binary> <tu> <asmdir> <arm> "<commit title>" fn… (the per-TU bank chain; refuses on __asm__/.word in a draft, any rtu non-MATCH, build rc≠0 or sha mismatch, and never commits on red — the tree is left for diagnosis) · backlog_near.sh (the 9 rows, re-runnable) · prompts/<fn>.txt (the 17 launched prompts) · gate/slate_main2.json + gate/gate_main2.log · r22_*.log / r22b_*.log / tools_health.log / build_*.log / bank_*.log / twin_rescan*.log.
• .run/P32/t3/: verdicts.jsonl now 48 rows (31 + 17; the S83 rows carry "session": "491895ad") · haiku/ 28 drafts (all banked) · opus/ 12 (2 banked S83: func_80015B6C, func_8002FDE8, func_801EF6D8 → 3 banked; 9 NEAR) · sonnet/ 4 (all banked) · reports/ (the 31 S82 reports + 3 Haiku reports) · pending_launch.txt (EMPTY — all 17 launched 10:57) · BRIEF.md (amended: §304 + exact-TU-spelling) · PROMPT_TEMPLATE.md · harvest_notes.md.
• Transcripts of the 17 S83 agents: ~/.claude/projects/-home-musashi-bfm-decomp/491895ad-3c84-4037-b04f-bf7e5ee16a0c/subagents/agent-<id>.jsonl (/tmp/claude-1000/-home-musashi-bfm-decomp/491895ad-3c84-4037-b04f-bf7e5ee16a0c/tasks/a*.output are symlinks; the b*.output there are this session's Bash jobs). All 17 verdicts are already in verdicts.jsonl; agent_verdicts.py <tasks>/a*.output rebuilds them; agent_reports.py saves the full reports.
• rtu_match: --split <sub> --source <binary> --c <draft> --asm-subdir <asmdir> --work <dir>; main adds --tu src/<sub>.c; verdict line MATCH (N ins) / DIFF / CC1 FAIL (compile-only: BLIND to link-time §304 and to cross-draft same-TU conflicts — the BUILD is the batch verdict). Build check: make build BINARY=<b> -j8, read the EXIT CODE (R53; a failed build leaves the OLD binary in place — the printed sha is then stale). gate_main: the ONLY way to gate main (slate JSON of {fn, draft(ABSOLUTE)}, --apply). Carves: for md_* island tables do NOT carve (§303 — the derive stage prints the pad spec in the build log); for overlays jtbl_carve.py <b> --func <fn> → make extract (rc) → build → pads_audit/interleave_check (R60).
• Gotchas that bit today: two drafts of one TU can each MATCH alone and fail together (spelling / §304 / a prototype a sibling bank introduced) — always rtu in the CURRENT TU right before splicing (bank.sh does) · backlog.py/corpus.stubs raise CorpusError … NO .s on disk while make extract-all is running (the instrument's environment, not a defect — R40) · census nins counts .s lines incl. rodata .words · agents read asm/ — never launch them while a make clean && make extract-all is in flight · the harness cap is 20 concurrent subagents · never cd inside a compound command · set -o pipefail + grep -c exits 1 on zero matches.

6. Carried context for T4 and T5

T4 (xHigh, no drafting) — the dead session's read-only pre-check (07:00Z) still stands: exclude_audit --assert-fresh: 7 entries, 7 WALL, 0 stale. Best drafts in their REAL TU: main:func_80011380 .run/m3/opus/func_80011380.c DIFF 6 (the ledger's .run/backlog_drafts/func_80011380.c is MISSING — R62 path fix) · main:func_80032A74 .run/S79w/opus/ func_80032A74.c CC1 FAIL (redefinition of struct Slot/Slot, conflicting types for A12/B12 vs src/800_shared.h:68/95 — §376 plumbing, re-probe before a final verdict, R40; the census prints closeness 1 for it from a stored masked diff) · main:func_80020DA4 .run/S79w/sonnet/func_80020DA4.c DIFF 2 · main:func_80039DEC both .run/S79w/{permuter,sonnet}/func_80039DEC.c CC1 FAIL (conflicting types for func_80039DEC — def-side declaration, §495: fix the TU's own decl first) · main:func_800391D4 .run/S79w/sonnet/func_800391D4.c CC1 FAIL (conflicting types for D_80073140 at t.c:3310 vs :2622) · ov_SC06_022:func_8017DF28 DIFF 2 · ov_SC03_105:func_801834A4 S71_gate14 variants DIFF 6 (ledger .run/backlog_drafts/func_801834A4.c MISSING). Deliverables: each of the 7 rows PROVED vs CANDIDATE with a current citation + an instrument-exonerated bounded attempt (R40 — the three CC1-FAIL rows get their §376 plumbing fixed byte-neutrally and are re-probed BEFORE any verdict), exclude_audit --write, backlog.py render, the wall table into this file, cookbook §501 if a verdict changes, decision-log, accelerators. A T3 idiom naming a wall's mechanism re-opens that row (one bounded attempt, permuter first). T5 (Max, Tier 1 — prompt R27; WAIT for gate 2): P7 checkbox walk over every task above; milestone demo = R22 fleet 218/218 · make tools-health OK · verbatim_check --strict 5==5 · the final census (tools/frontier_classify.py --json) · the wall ledger · parked-5 dispositions via make audit-disc (UNCLAIMED 0 of 220 at T2c) · make report all three metrics + main 143dbb89… WITH and WITHOUT the SDK object dirs (.run/obj40, .run/obj42; fresh-extract fallback) · corrected denominators (main game-code 41,556 — the 22-ins Ghidra-boundary gap; the fleet denominator grew by the five modules' ~2,600 ins at T2b). Then PhaseEnd_Phase32.md (+ Roadmap delta: P33 = verify + flip; + Plain-English Recap; + the rules table incl. R64) → append the P32 synopsis + rules to phase-ends/DIGEST.md (step 3b) → git mv CURRENT_PHASE.md phase-ends/logs/Phase32.md (R19) → R23 MCP stop if it ran → leave both uncommitted for Drew (R6) → 🛑.

7. Environment

Fleet 218; shas: main 143dbb89f34491258bbc27810d0a12ec8b43a8dd · resident 8e17e02ff8954d07c979449198f7e1645046b353 · md_SC03_053 c0848f308342e499e2bd8c2db86f6225a29e1592 · md_SC03_054 06bd73dfbdb10748adcbac7093ead5d76131206a · md_MAIN_007 2ff702b605ab5cfc18474c464c4c07e5f8ffd48c · md_MAIN_009 d270f695b793b5c03db159b7aabcc066daa87eda · md_SC03_056 bc768a6b7baa911e0a0cb9175b33550851d6c1eb (config/check.<b>.sha is authoritative). Toolchain unchanged (SETUP §5.4). The SessionStart hook launches a headless Ghidra MCP (.run/ghidra-mcp.log); this session's MCP failed to connect and nothing needed it — T4/T5 do not either (R23: never stage ghidra/). Harness: 20 concurrent Agent-tool subagents max. Metrics: regenerate with make report at the T3 close (218 binaries; last published = P31 close, 213 binaries: instr 100.0% · distinct 99.9% · fn 99.99%).

8. Standing reminders (the rules that bite now)

Every bank commits before the next command that can touch src/ (R42) · count banks from the SOURCE and quote the fleet's green count at every close (R58) · verify a build from its exit code (R53) · a masked/permuter score is not a closeness until its diff is read (R63) · an __asm__ body is a verbatim, not a bank (R62) · fix the TU by a byte-neutral plumbing commit BEFORE any gate (S77 law) · re-verify in the CURRENT TU before splicing (§500-F) · rules check every 4 tasks (P6) · harvest idioms into the cookbook before the next drafting step (R16/R30) · no Co-Authored-By (R5) · never stage ghidra/ (R23).

Plain-English recap (R18). This session banked twenty-eight functions in about an hour: first the ten exact matches the overnight agents had already produced (after re-checking each one byte-for-byte), then seventeen small functions drafted by seventeen cheap agents launched together, every one of which came back exact on the first try. Three of the game's code modules are now fully rebuilt from C, and the whole 218-file set still rebuilds identically. What is left is sixteen functions: seven long-standing "walls" that get their final written verdicts next, and nine near-misses that are a few instructions off, which get one bounded round with the brute-force polisher and one careful second look before the phase closes.