BFM-decomp

distill

docs(cookbook): 268-269 from the ax/bm distill batch (82 candidates)

2026-08-24 14:09:13 -06:00

gcc-2.7.2-map

docs(phase-29): regalloc.md §H — the swap oracle's two unstated PRECONDITIONS

2026-07-28 15:02:51 -06:00

history

feat(phase-22): cheap-tier A/B experiment harness + findings

2026-06-29 12:52:39 -06:00

tool-designs

docs(phase-31): T4 DONE — routing rule (<=50: Sonnet+DeepSeek parallel, Opus residue; 51-120: Sonnet, Opus escalation; >120: Opus; haiku retired; M-extend-tell -> wall); frontier-s61 Addendum 5; judge artifacts (P31 S62)

2026-08-26 18:32:51 -06:00

accelerators.md

docs(phase-32): T4c harvest + checkpoint — the frontier is EMPTY (census 0 stubs / 0 ins); cookbook §501-Q (the combine self-update ghost slot) + §501-R (a hoisted invariant read three times: inline temps merged by combine_movables, u16 as a cse firewall, the allocation order IS the callee-saved bank; every pin came off), accelerators (15)(16), decision-log P32 S85 T4c, tools/cc1_dumps_tu.sh + tools/alloc_table.py promoted (+ SETUP rows, cc1_dumps.sh -dL), wave_exclude 0 entries, backlog 1 legacy row, .run/P32/t4e (56 variants, 18+8 reproducers, NOTES, banks, slates); 🛑 checkpoint refreshed as the T5 seed (the fleet R22 + report chain runs in the background — its logs are recorded by the next session)

2026-09-06 16:13:19 -06:00

actor-struct.md

feat(phase-17): T4+T5 — actor struct recovered (comprehension) but byte-neutral for matching

2026-06-19 13:03:14 -06:00

automation-runbook.md

docs(p31 s67): wave-playbook (the CURRENT pipeline) + seed_ref fix + harvest §333-§338

2026-08-31 13:48:56 -06:00

autopsy.md

feat(phase-30): T0e2+T0f — autopsy refreshed (108 stranded MATCHes); frontier regenerated; the three populations PINNED (docs/frontier-p30.md)

2026-07-30 16:45:09 -06:00

backlog.md

feat(phase-33): A2 reporting instruments — make sig-main rewritten (tools/main_seed_ends.py: main's game-code boundaries DERIVED from the link map + objects, tiling asserted; 809 fns / 45,150 ins), progress.py weighs main by the build-derived sig and EXITS with no sig (R32), the digest's oracle clause derived live (0 phantom / 0 truncated / 0 pad-tail), backlog.linked_closed retires the LINKED-range legacy row (0 open), dup_report on the derived sig; CORRECTED main denominator 45,150 (Ghidra's flow boundaries left 3,628 words of game code unowned) — fleet instr 13,492,113 / distinct 5,820,205, all 100%; make report BINARY=main EXIT=0, audit-digest OK; SETUP §6.8 + P33 A2 rows (R21)

2026-09-06 18:46:52 -06:00

burndown.jsonl

chore: regenerated burndown

2026-09-02 16:45:33 -06:00

calibration.md

feat(phase-30 S39/S4): re-gate probe — A10 broadly stands; 4 banked from the reverted overlays (+146 ins)

2026-08-04 22:42:54 -06:00

canonical-decl-reconcile-design.md

docs(phase-24): T7b — distill reconcile_decls (cookbook §33, SETUP, design-doc RESOLVED); R14-correct §32 wall

2026-07-03 17:59:19 -06:00

community-matching-model-plan.md

feat(phase-23): broad v3 harvest (48 banks) + final propagate sweep; docs

2026-06-30 13:54:53 -06:00

concurrency-design.md

docs(phase-30 S45 p7): F1 confirmed live (cookbook 156) + the cheap-tier size cliff (157) + wave_snapshot

2026-08-07 18:04:26 -06:00

cookbook-index.md

docs(phase-32): cookbook §501 — a lever that measures worse may be a cascade (func_800391D4: the explicit promotion flipped loop.c's insn_count threshold; pads 7→9 closed it); accelerators (9); T4b row

2026-09-05 13:16:50 -06:00

debug-menu-list.txt

feat(phase-30 S45 L3-p3): SC02/9 = the Steam Knight boss module — decoded, captured, retro-verified, onboarded; parked = 5

2026-08-07 15:07:55 -06:00

decision-log.md

docs(phase-32): T4c harvest + checkpoint — the frontier is EMPTY (census 0 stubs / 0 ins); cookbook §501-Q (the combine self-update ghost slot) + §501-R (a hoisted invariant read three times: inline temps merged by combine_movables, u16 as a cse firewall, the allocation order IS the callee-saved bank; every pin came off), accelerators (15)(16), decision-log P32 S85 T4c, tools/cc1_dumps_tu.sh + tools/alloc_table.py promoted (+ SETUP rows, cc1_dumps.sh -dL), wave_exclude 0 entries, backlog 1 legacy row, .run/P32/t4e (56 variants, 18+8 reproducers, NOTES, banks, slates); 🛑 checkpoint refreshed as the T5 seed (the fleet R22 + report chain runs in the background — its logs are recorded by the next session)

2026-09-06 16:13:19 -06:00

difficulty.md

chore(phase-32): CLOSE — the frontier is empty: 0 stubs across 218 binaries, fleet 100.00 / 100.0 / 100.0, main 2,091/2,091 (v1.31.0)

2026-09-06 16:29:00 -06:00

difficulty.ov_SC01_005.md

docs(phase-13): T6 docs — cookbook §13 overlay runbook + SETUP/gen2-roadmap + digests

2026-06-16 17:56:38 -06:00

difficulty.ov_SC01_006.md

docs(phase-13): T6 docs — cookbook §13 overlay runbook + SETUP/gen2-roadmap + digests

2026-06-16 17:56:38 -06:00

difficulty.ov_SC01_077.md

feat(phase-19): T2 recovery tooling — canon_resident_calls + no-proto caller fix (3 matches)

2026-06-20 17:12:00 -06:00

difficulty.ov_SC03_001.md

docs(phase-13): T6 docs — cookbook §13 overlay runbook + SETUP/gen2-roadmap + digests

2026-06-16 17:56:38 -06:00

difficulty.resident.md

feat(phase-12): T3a — extern-conflict recovery, +20 (REAL 102->123, 85.6%)

2026-06-16 14:05:20 -06:00

disc-completeness.md

docs(phase-32): T2c CLOSE — 20 free banks, R22 218/218 rc 0, census 54 stubs / 6,376 ins; disc-completeness P32 section (parked ledger EMPTY, audit-disc UNCLAIMED 0 of 220); regenerated fleet/disc/dup digests; T2d not needed

2026-09-05 00:49:07 -06:00

disc-ledger.md

docs(phase-32): T2c CLOSE — 20 free banks, R22 218/218 rc 0, census 54 stubs / 6,376 ins; disc-completeness P32 section (parked ledger EMPTY, audit-disc UNCLAIMED 0 of 220); regenerated fleet/disc/dup digests; T2d not needed

2026-09-05 00:49:07 -06:00

duplicates.cross.md

feat(phase-33): A2 reporting instruments — make sig-main rewritten (tools/main_seed_ends.py: main's game-code boundaries DERIVED from the link map + objects, tiling asserted; 809 fns / 45,150 ins), progress.py weighs main by the build-derived sig and EXITS with no sig (R32), the digest's oracle clause derived live (0 phantom / 0 truncated / 0 pad-tail), backlog.linked_closed retires the LINKED-range legacy row (0 open), dup_report on the derived sig; CORRECTED main denominator 45,150 (Ghidra's flow boundaries left 3,628 words of game code unowned) — fleet instr 13,492,113 / distinct 5,820,205, all 100%; make report BINARY=main EXIT=0, audit-digest OK; SETUP §6.8 + P33 A2 rows (R21)

2026-09-06 18:46:52 -06:00

duplicates.md

feat(phase-33): A2 reporting instruments — make sig-main rewritten (tools/main_seed_ends.py: main's game-code boundaries DERIVED from the link map + objects, tiling asserted; 809 fns / 45,150 ins), progress.py weighs main by the build-derived sig and EXITS with no sig (R32), the digest's oracle clause derived live (0 phantom / 0 truncated / 0 pad-tail), backlog.linked_closed retires the LINKED-range legacy row (0 open), dup_report on the derived sig; CORRECTED main denominator 45,150 (Ghidra's flow boundaries left 3,628 words of game code unowned) — fleet instr 13,492,113 / distinct 5,820,205, all 100%; make report BINARY=main EXIT=0, audit-digest OK; SETUP §6.8 + P33 A2 rows (R21)

2026-09-06 18:46:52 -06:00

duplicates.ov_SC01_077.md

feat(phase-19): T2 recovery tooling — canon_resident_calls + no-proto caller fix (3 matches)

2026-06-20 17:12:00 -06:00

duplicates.resident.md

feat(phase-10): T4 — resident Ghidra program (2nd binary) + symbol seed + sig

2026-06-15 23:25:00 -06:00

effort-map.md

feat(phase-13): T7 — effort experiment (Max vs Ultracode yield = noise) + bank 14

2026-06-16 17:50:31 -06:00

exemplar_curriculum.md

feat(phase-20): T2 — exemplar_miner routes 835 residuals (WAVE 472 / PINS 45 reach-134 / STUB 90)

2026-06-20 23:55:51 -06:00

fable5-strategy-review-2026-07.md

docs(phase-23): Fable5 strategy review — scope-vs-reality scorecard, wall re-tests, resource map

2026-07-01 17:25:47 -06:00

family-cousins.md

chore(docs): sweep auto-rendered ledgers at S61 close

2026-08-26 12:56:10 -06:00

family-endgame-megaplan.md

docs(roadmap): adopt the Road-to-100 endgame roadmap (Phase 27+) — supersedes the family-endgame megaplan

2026-07-15 15:49:45 -06:00

family-hseq.md

fix(phase-33): A4 family map carries its own coverage — family_hseq.json records the 217 scanned binaries + open_instances (0), audit_binaries CHECK 4 reads that denominator instead of inferring it from (now empty) families; regenerating the map had turned the P32-close 6-binary warn into 217 'missing' from a COMPLETE map (R35); controls: 0 [warn] now, the old format warns 'predates the coverage field'; docs/family-hseq.md regenerated (0 families); SETUP P33 A4 (R21)

2026-09-06 18:55:21 -06:00

family-manifest.md

docs(phase-29): regen family-manifest after session-4 banks (frontier snapshot)

2026-07-19 00:36:51 -06:00

formats.md

docs(phase-12): T5 — save trace findings (memcard path + directory confirmed; deep data deferred)

2026-06-16 14:33:54 -06:00

frontier-atlas.md

docs(phase-31): frontier-analysis-s61 — the post-ox finish-plan ground truth (2,520 open instances fully classed, zero unassigned; wall = 86 fns; done-denominator reconciled; engines measured per class)

2026-08-26 12:45:16 -06:00

frontier-p30.md

feat(phase-30): T2 — rollout_o0.py (generalized o0b driver) + x1 probe verdict: append route REFUTED for the remaining cluster (R14 correction)

2026-07-30 17:05:50 -06:00

frontier-p32.md

docs(phase-31): S80 — docs/frontier-p32.md, the Phase-32 jumping-off document: the last 21 functions by blocker (each with its instrument's verdict, mechanism citation, best-draft path and route — the two NON-CONTIGUOUS carves dry-run/refusal detail, the RELOC-ONLY twin, the leading-island case, the 8 pinned walls, the 3 near plateaus, the 7 far drafts), the five parked disc payloads, the routes/instruments a fresh session must know, and the proposed Phase-32 shape with its kill gate

2026-09-04 22:20:20 -06:00

frontier-s39.md

docs(phase-30 S39): burn-down for T5 prep — with the -0.14pp last row annotated as a denominator correction

2026-08-05 00:50:01 -06:00

gen2-mips-matching-model.md

docs(phase-25): T4 CLOSEOUT — v4 retrain FAILED (negative), keep v3; T5 handoff

2026-07-08 18:36:50 -06:00

gen2-roadmap.md

docs(phase-15): close — PhaseEnd v1.14.0 (fleet 54.48%), continuation backlog -> Phase 16

2026-06-18 21:01:38 -06:00

gen3-parking-lot.md

feat(phase-25): T5b batch-1 COMPLETE — continuation +541 fns (fleet 72.02->72.18%)

2026-07-09 15:47:30 -06:00

generic-decomp-package.md

feat(seed_ref): the CONTAINED tier + docs for the twin ladder (§390/§391, accelerator #18 )

2026-09-01 13:06:27 -06:00

hand-matching-process.md

fix(phase-26a): A9d — retire the dead Phase-17 canonical-sig chain (R33)

2026-07-14 21:21:02 -06:00

hindsight-study.md

docs(phase-29): hindsight study — optimal decomp ordering + offline-automatic endgame (§7)

2026-07-20 19:24:17 -06:00

idxtab-map.md

feat(phase-30 S46-2): the master IDXTAB/DESTPTR load map — and the tracker blind spot that hid it

2026-08-07 22:09:05 -06:00

matching-cookbook.md

docs(phase-32): T4c harvest + checkpoint — the frontier is EMPTY (census 0 stubs / 0 ins); cookbook §501-Q (the combine self-update ghost slot) + §501-R (a hoisted invariant read three times: inline temps merged by combine_movables, u16 as a cse firewall, the allocation order IS the callee-saved bank; every pin came off), accelerators (15)(16), decision-log P32 S85 T4c, tools/cc1_dumps_tu.sh + tools/alloc_table.py promoted (+ SETUP rows, cc1_dumps.sh -dL), wave_exclude 0 entries, backlog 1 legacy row, .run/P32/t4e (56 variants, 18+8 reproducers, NOTES, banks, slates); 🛑 checkpoint refreshed as the T5 seed (the fleet R22 + report chain runs in the background — its logs are recorded by the next session)

2026-09-06 16:13:19 -06:00

memory-map.md

feat(phase-32): T2b (4) — SC03/56 ONBOARDED as md_SC03_056 @0x801CBB50 (ov_SC03_002's DESTPTR), byte-identical bc768a6b; ALL FIVE parked payloads are now binaries (fleet 213 -> 218); evidence tool v2

2026-09-04 23:52:58 -06:00

next-session-triage-ladder.md

docs: mark the triage-ladder spec BUILT with its three corrections; accelerator #16

2026-09-01 03:49:21 -06:00

phase25-frontier-map.md

docs(phase-25): frontier map — rtu_match-measured classification of the 42 draftable + 235 matched-free families

2026-07-10 17:54:01 -06:00

phase25-t6-curriculum.md

feat(phase-25): T6 — Fable5 crack curriculum: def-side wall ~71% tool-shaped; 44/62 probe-BANKABLE via canon_sig_reconcile v3.1; x134 sweep law proven

2026-07-09 23:39:14 -06:00

phase25-t6-fable-brief.md

chore(phase-25): T5b batch-3 (_o0 measured) + T6 Fable5 brief + match_one --o0 (fleet 72.29%, handoff)

2026-07-09 22:45:56 -06:00

phase26-triage.md

docs(phase-26): finalize Task-6 triage (119 agents) + banking caveat; pause before Task 8

2026-07-11 23:12:53 -06:00

portable-decomp-workflow.md

docs: seed the portable decomp workflow — the part of this project that transfers

2026-08-24 15:46:01 -06:00

progress.fleet.md

feat(phase-33): A2 reporting instruments — make sig-main rewritten (tools/main_seed_ends.py: main's game-code boundaries DERIVED from the link map + objects, tiling asserted; 809 fns / 45,150 ins), progress.py weighs main by the build-derived sig and EXITS with no sig (R32), the digest's oracle clause derived live (0 phantom / 0 truncated / 0 pad-tail), backlog.linked_closed retires the LINKED-range legacy row (0 open), dup_report on the derived sig; CORRECTED main denominator 45,150 (Ghidra's flow boundaries left 3,628 words of game code unowned) — fleet instr 13,492,113 / distinct 5,820,205, all 100%; make report BINARY=main EXIT=0, audit-digest OK; SETUP §6.8 + P33 A2 rows (R21)

2026-09-06 18:46:52 -06:00

progress.md

chore(phase-32): CLOSE — the frontier is empty: 0 stubs across 218 binaries, fleet 100.00 / 100.0 / 100.0, main 2,091/2,091 (v1.31.0)

2026-09-06 16:29:00 -06:00

progress.ov_SC01_005.md

docs(phase-13): T6 docs — cookbook §13 overlay runbook + SETUP/gen2-roadmap + digests

2026-06-16 17:56:38 -06:00

progress.ov_SC01_006.md

docs(phase-13): T6 docs — cookbook §13 overlay runbook + SETUP/gen2-roadmap + digests

2026-06-16 17:56:38 -06:00

progress.ov_SC01_077.md

feat(phase-19): T2 recovery tooling — canon_resident_calls + no-proto caller fix (3 matches)

2026-06-20 17:12:00 -06:00

progress.ov_SC03_001.md

docs(phase-13): T6 docs — cookbook §13 overlay runbook + SETUP/gen2-roadmap + digests

2026-06-16 17:56:38 -06:00

progress.resident.md

feat(phase-28 T5): resident 21 -> 14 stubs (7 banked, 90.34%) + fix match_one's fake isolation

2026-07-16 01:41:57 -06:00

proto-correspondence.md

feat(phase-3.5): prototype research spike — go/no-go (protos don't accelerate Gen1 matching)

2026-06-14 03:19:23 -06:00

psyq-sample-idioms.md

feat(phase-31 T2): reference expansion — PsyQ SDK sample source extracted + gcc-2.7.2 completed

2026-08-14 17:44:32 -06:00

psyq-worklist.md

feat(phase-31): S79 #5 — the libpad 4.2.1 + libapi 4.2 band and the apicard region LINKED from real objects: 13 stubs + 4 TUs + the reorder island gone; main 16 stubs, fleet 38

2026-09-04 17:56:31 -06:00

recovery-queue-s65.md

docs: recovery queue — the step-2 lane I skipped for 12 waves, classified and QUEUED

2026-08-29 20:29:34 -06:00

resident-dossier.md

docs(phase-28 T5b): the resident wall dossier — 14 remaining, classed and byte-grounded

2026-07-16 01:44:11 -06:00

roadmap-to-100.md

docs(phase-30 S45 II.3): metrics re-baseline + roadmap contract delta + decision-log (R31)

2026-08-06 13:36:34 -06:00

second-oracle.md

feat(oracle): main gets its independent second oracle — contract §1.3 closed

2026-09-03 21:58:17 -06:00

SETUP.md

feat(phase-33): B1 make disc-extract — the rom->decoder step in the build: extract.py --expect-manifest (compare against the committed oracle, never write it; mismatch -> .run/extract/ + diffs, exit 1) + --allow-missing-audio (explicit PARTIAL for Track-1-only dumps); disc-extract = probe (0.7 s no-op) -> disc presence -> redump SHA1/CRC32 -> extract+compare -> verify (15.7 s full); extract/extract-all call it; check-env WARN-on-absent EXE + oracle self-consistency; help rewritten; the 4 splat preset headers TRACKED (clean keeps them); .gitignore re-tightened to H1 (dumps/*.bin, ghidra/, tools/psyq/, session archive/, ghidra-ext zips, brave.exe; EXE re-include dropped); controls: no-disc exit 2, truncated disc FAILs with the oracle untouched, regenerated tree byte-identical to the previous (1,801 files)

2026-09-06 19:02:34 -06:00

sister-decomps.md

feat(phase-23): broad v3 harvest (48 banks) + final propagate sweep; docs

2026-06-30 13:54:53 -06:00

struct-core-pivot.md

feat(phase-17): T4+T5 — actor struct recovered (comprehension) but byte-neutral for matching

2026-06-19 13:03:14 -06:00

tooling-audit.md

docs(phase-30 S44 I.0): the static loader routing table + the full tool audit — knowledge captured

2026-08-06 10:52:07 -06:00

wall-taxonomy.md

feat(phase-17): T3 — bank 42 sig_unify matches + propagate x134 (fleet 54.48%->55.00%); fix taxonomy classifier

2026-06-19 12:13:38 -06:00

wave-metrics.md

chore(phase-31): S54/S55 fresh-session checkpoint — 547 banked, fleet 96.0/91.7/97.07, wave Z staged

2026-08-18 12:41:37 -06:00

wave-playbook.md

docs(phase-32): T3 checkpoint — the 31 agents' FULL final reports saved to .run/P32/t3/reports/ (tools/agent_reports.py, NEW) + tools/transcript_dump.py (NEW, the dead-session reader); SETUP rows, playbook addendum-2, 🛑 block §3/§6 point at them

2026-09-05 10:11:43 -06:00

worklist.md

docs(phase-30): T6 approved — P30 stays open; the measured-frontier continuation (S1-S5)

2026-08-01 08:20:41 -06:00