Files
aetherbound-guild/docs/goals/PHASE9_0R7_INDIVIDUAL_GROWTH_GOAL.md
T

6.1 KiB

Phase 9.0R7 Individual Growth Goal

Owner Decision And Role

On 2026-08-13 the Owner reviewed P9.0R6 and asked to continue with a larger playable delivery.

  • Role: development task.
  • Repository: aetherbound_guild only.
  • Baseline: clean/upstream P9.0R6 closure 89ed89b2a5b45ee5ab10c989736f2d8c65ea88dd.
  • Preserved P9.0R6 product source: 3620e5a7cf6b762fefc3eb92d72b465c8bb209eb.
  • Active contract: docs/runtime/PHASE9_0R7_INDIVIDUAL_GROWTH_CONTRACT.md.
  • Active ledger: docs/reviews/PHASE9_GODOT_VERTICAL_SLICE_LEDGER.md.

Player Hypothesis

The repeated expedition loop becomes growth rather than a sequence of route receipts when one earned decision changes one individual Recruit and the next battle visibly depends on which role received that growth.

The owned continuation is:

P9.0R6 second-cycle result -> individual training -> review/commit -> fifth automatic battle -> third-cycle result

Owned Slice

  1. Preserve the complete P9.0R6 Title-to-second-cycle-result path unchanged.
  2. Award one run-local Veteran Mark / 历练印记 at the verified R6 result. It is a one-use decision receipt, not a new currency or inventory system.
  3. Show every actual living Company member at Lv.1. The player selects one stable slot (front, rear, or present added), can revise it, reviews the exact role effect, and commits once.
  4. Only that member becomes Lv.2. There is no shared Company level, XP pool, promotion, skill tree or transfer.
  5. Run one fifth automatic battle against an existing Riftblade-family route guard. It applies one front crush and one original-rear split so front, Farstring and Kindhand growth solve different observed problems.
  6. Result leads with trained member -> changed event -> outcome, then exact front/rear HP, time, one-time +8 Coin, Company and member levels.

Growth Counterfactuals

  • front Lv.2: reduces the front crush by another 20. Retained EQ-002 still reduces the base -45 crush to -30, so both together leave front HP 90; without Frontpin, trained front ends at 75.
  • Farstring Lv.2: performs an early shot and changes declared battle time from 7.40s to 6.40s; HP outcomes do not improve.
  • Kindhand Lv.2: restores the split rear from 60 to 85 instead of the normal 72; battle time does not improve.

The three outcomes answer survival, tempo and recovery respectively. None is labeled recommended. Every actual member remains selectable; a two-person Company has two choices and a full line has three.

Fast H5 Gate

  • Focused domain coverage across all R6 build/route/original-rear states and every legal training slot, including invalid slot, revise, cancel, once-only commit/reward and restart cleanup.
  • Parsed ScreenTouch from the real P9.0R6 second-cycle result through training, revise, review, cancel, fifth battle and third-cycle result with fallback=0 direct=0.
  • P9.0R6/P9.0R5/P9.0R4/P9.0R3 regressions.
  • Normal-render 844x390 Simplified-Chinese training, review, three battle counterfactuals and three causal result states, with every action at least 44px and no scrolling.
  • Fixed-source runtime-only Web export, PCK scope check, local and remote normal-GPU Chrome Title launch, and a new Gallery item.

Frozen

  • P9.0R3 through P9.0R6 values, presentation and complete flow.
  • Character/background/VFX/animation source assets and paid generation.
  • Authoritative XP curves, levels 3..10, promotion, skills, Traits, Market, permanent inventory, Party ordering, Reserve, casualty/recovery and save.
  • Chapter/Region content expansion, authoritative combat, device packaging, TestFlight, stores and release.

Acceptance

  • Training opens only from a verified R6 result and starts unselected.
  • Only actual Company slots render; invalid or absent slots cannot commit.
  • Selection is reversible, review can cancel, and commit consumes the one mark exactly once.
  • Exactly one member reaches Lv.2; all others remain Lv.1.
  • Front, Farstring and Kindhand produce the declared distinct HP/time effects with no hidden reward change.
  • Fifth-battle reward adds 8 Coin exactly once.
  • Third-cycle result names the trained role and changed event before totals.
  • Restart/Home clear all three cycles; P9.0R6 regression remains unchanged.

Aetherbound Guild P9.0R7 individual growth H5 preview ready

or

Aetherbound Guild P9.0R7 redesign required: <earliest failed player boundary>

Fast H5 Result

Product source b7d50af403d3fe5d3fcc92943bc21d1d1748ae3d is clean, upstream and completes the owned continuation:

R6 second-cycle result -> individual training -> review/commit -> fifth battle -> third-cycle result

The fast gate passes:

  • ABG_P9_0R7_INDIVIDUAL_GROWTH_OK for two- and three-person legal member sets, unselected/revision/review-cancel, once-only training/reward, exactly one Lv.2, all front/Farstring/Kindhand counterfactuals, restart cleanup, front HP 90/75/70/55, rear HP 85/72/60 and time 7.40/6.40;
  • ABG_P9_0R7_GROWTH_TOUCH_OK touch=8 fallback=0 direct=0 from a real R6 result through training revision, review cancel, commit, battle and result;
  • the declared P9.0R3-R6 regressions;
  • eleven normal-render 844x390 Simplified-Chinese states with controls at least 44px, no clipping or overlap, and visibly distinct mitigation, early-shot and stronger-recovery feedback; and
  • a fixed-source runtime-only PCK with tests/evidence/build absent, nine delivery payloads verifying against export.sha256, local normal-GPU Chrome Title, and remote signed normal-GPU Chrome Title with browser warning/error logs empty.

The ignored delivery is runtime/build/web-preview/b7d50af-r7-fast/. export.sha256 has SHA-256 6a2b04e823da55af21b658d1d96bf3faf969849496cdabe077770ca0fa55066e. The new private Artifact Gallery item is REs3L6178z7p; it does not overwrite R6. Remote cold start remains slow because the unchanged roughly 53 MB Web bundle is delivered uncompressed. This is a fast preview only. Owner growth- choice comprehension, fun/continuation, listening, physical device, packaging, stores and release remain open.