Files
aetherbound-guild/CURRENT_CHECKPOINT.md
T

111 lines
6.0 KiB
Markdown

# Current Checkpoint
Active Goal: `docs/goals/PHASE8_EXPERIENCE_PROTOTYPE_GOAL.md`
Active contract: `docs/presentation/PHASE8_EXPERIENCE_PROTOTYPE_CONTRACT.md`
Active ledger: `docs/reviews/PHASE8_EXPERIENCE_PROTOTYPE_LEDGER.md`
Baseline: `b3968b0f41f841efb0f699d8fb04a9bcb958182a`
Accepted P8.1 source: `0b4a64fe549af0ff965e654aa4f5fb7f2ff054b4`
P8.2 candidate source: `86a8a1adec39de40b4adfee6b01033407e3e3421`
Accepted P8.2 closure: `d308f8bc14b043ed133ce2d99cbce461aecec81f`
Rejected P8.3 source: `5f0cb2695ba149b1efcab27a882ade00a275c6ac`
P8.3 documentation closure: `fab5eebf4260ab3da42dee8abe63dd7b7588789b`
Stage: `P8_3R1_DYNAMIC_BATTLE_ACTIVE`
The Phase 6 battle, VFX and equipment direction is Owner-accepted. Phase 7
turns the existing 50-screen design inventory into an individually viewable,
Chinese-first, production-oriented landscape HTML atlas. Runtime, engine,
audio, bulk asset production, device, package and release work remain frozen.
Strongest evidence: fixed-source static design validation, exact
asset-provenance checks, browser interaction and responsive-layout
verification, offline self-contained bundle verification, and agent visual
review. The Phase 7.1 candidate evidence is ignored local evidence at
`prototype/gallery/evidence/r7-1-candidate-e23b00b-attempt01/`. Its browser log
SHA-256 is
`b735e2218321b4452e3ea2a49b0e03d6c2f568cd218fb15f340d9fe635453441`;
the 23-row inventory SHA-256 is
`a169dbdc4d03eb00a6ffffc447d035cf5cb0a423422af9060d4eaff81d0b2706`
and verifies 23/23 from its evidence root. The self-contained HTML bundle
SHA-256 is
`64dbe54f94e66d0bff6bcdf17f5852c5c77c8b56b25eef8f30068ab9198f8176`.
Phase 7.1 remains the accepted automated design baseline: every Recruit has one
active equipment slot and individual run-scoped level/XP capped at level 10.
The Owner has authorized Phase 8 as five separately playable experience
packages. On 2026-08-12 the Owner reviewed the published P8.1 Contract and
receipt flow and explicitly accepted it. P8.2 Market Preparation is now the
only active package; P8.3-P8.5 and Godot/runtime development remain frozen.
The authoritative ignored local evidence is at
`prototype/gallery/evidence/p8-1-candidate-0b4a64f-attempt01/`. Its nine-file
inventory verifies 9/9 with SHA-256
`b076e5990aadb5e2b72ed979e27403a294c17b342c9bbadbad63b460c7fd60f3`.
The P8.1 browser log SHA-256 is
`c7e9a7278cb99881b51715bc25e3bd3112293ad282b63bc8449c1150653f6729`;
the Phase 7 regression log SHA-256 is
`af5eb0616e7dc1061f702d392952419701a93b6452475d626fac490d5c8f3aae`;
the self-contained bundle SHA-256 is
`e5895441d2835a22e9f94d6615c95be7180b7df6983c4611b499efc730feb915`.
P8.2 uses the authority price formula rather than the obsolete Phase 7 sample:
each level-1 base Recruit costs 14 Coin, so two separate hire receipts move the
opening balance `52 -> 38 -> 24` and Company `2 -> 3 -> 4` while Line capacity
remains four. It then owns one equipment transaction and one tap-select/tap-
destination order change, stopping before Risk. Fixed clean/upstream source
`86a8a1adec39de40b4adfee6b01033407e3e3421` passes the complete P8.2 browser
chain, P8.1 regression, complete Phase 7 browser regression, design validation,
anti-dominance, single-equipment/individual-level and asset-provenance gates.
The authoritative ignored local evidence is at
`prototype/gallery/evidence/p8-2-candidate-86a8a1a-attempt01/`. Its 42-row
inventory verifies 42/42 with SHA-256
`d41bed33f982b6504f91aaa6830d2ce09422d9b5dbbd70cfdf7c46d08f5b053e`.
The P8.2 browser log SHA-256 is
`5eab37da40298740b76f598701b983a53584abef4f61e1f0c43bf3210a69d0a6`;
the P8.1 regression log SHA-256 is
`564964d31a0bb1cf4c2f8ccb3611c8c94863d4f331d67e238851112773016f20`;
the Phase 7 regression log SHA-256 is
`55616bf1325da8bce24fbe6d52adf002382ee37cd4ce0e7e780b0f0696b53d2f`;
and the self-contained bundle SHA-256 is
`96751234e01a2c3c66fc4065be2db20517efd50abf0ced18075676ec42080251`.
On 2026-08-12 the Owner reviewed the published P8.2 experience and replied
`OK`. P8.2 comprehension and visual direction are accepted for progression.
P8.3 is the only active package: compare three risks, commit one, observe a
bounded automatic battle, pause/change speed and inspect one causal event.
P8.4-P8.5, authoritative battle simulation, Godot/runtime, audio, device,
packaging and release remain frozen.
Fixed clean/upstream P8.3 source
`5f0cb2695ba149b1efcab27a882ade00a275c6ac` passes the three-risk selection,
single commit, frozen opening, six-event automatic presentation, pause/speed,
read-only causal inspection and result-withheld terminal gates. It also passes
P8.1/P8.2 and complete Phase 7 browser regressions plus unchanged design,
anti-dominance, single-equipment/individual-level and asset-provenance gates.
The authoritative ignored local evidence is at
`prototype/gallery/evidence/p8-3-candidate-5f0cb26-attempt01/`. Its 53-row
inventory verifies 53/53 with SHA-256
`3ad48cfa135a29f01f75ed69dc07f6a2037fbf984c3d94b7cae21da2199af1c8`.
The P8.3 browser log SHA-256 is
`f34ec892e12dfd5022182de2b5f3f68687a77e1066a78aa2fe8e81bdd3ea7e21`;
the P8.2 regression log SHA-256 is
`3139beb66a7f265a40c29e02518ad63ffa802bb10f9ab819cb62cf7e39060747`;
the P8.1 regression log SHA-256 is
`c38118d8b66bb1270e625586f5d9ae77af85ae60ae0c1ab88631a2f764fc6a8b`;
the Phase 7 regression log SHA-256 is
`b2b8329301b944af8a9301da4a4c493007730f66a310e6942de67ba00a3343c8`;
and the self-contained bundle SHA-256 is
`5989cb87c5e422d26bd1cdc5d2e74db4177090b32540d1cb53b77560bec943e9`.
Owner review rejected the prior P8.3 battle presentation because static unit
cards and the central event explainer made it read as a timeline board rather
than an animated horizontal automatic battle. P8.3R1 is bounded to rebuilding
that battle presentation around one continuous ground line with visible
advance, attack, projectile, intercept/Guard, hit and sustain motion. The
accepted P8.1/P8.2 behavior and the P8.3 risk/commit/event/result boundaries
remain unchanged; P8.4 remains locked.
Shortest next action: implement and verify the P8.3R1 dynamic horizontal battle
presentation without entering result/recovery.