717 lines
36 KiB
Markdown
717 lines
36 KiB
Markdown
# Aetherbound Guild - Game Design
|
|
|
|
Product promise, audience, complete game scope, core loop, duration, completion and originality boundary.
|
|
|
|
## Owner Review / Owner 审核
|
|
|
|
先确认游戏幻想、目标玩家、核心循环、单局/全局时长、完成条件和原创边界。这里回答“这是什么游戏、为什么值得玩、最终做到哪里”。
|
|
|
|
本文件是当前六份审核文档之一。历史拆分文件仅保留在 Git 历史中;文内规则、表格和 ID 仍是完整开发依据。
|
|
|
|
## Current Consolidated Project Contract
|
|
|
|
> Status: `OWNER REVIEW REQUIRED / P0 PROJECT CONSOLIDATION`
|
|
> Purpose: one positive product contract that connects product strategy,
|
|
> systems, pages, presentation and ordered development. Detailed authority
|
|
> remains in Sections 2-16 and the companion documents below.
|
|
|
|
### Product Identity
|
|
|
|
Aetherbound Guild is a landscape Japanese 2D hand-drawn fantasy incremental
|
|
management game. The player grows a persistent Expedition Company made from
|
|
generated Recruits, prepares a horizontal Party line, commits bounded Contracts,
|
|
observes deterministic automatic battles and adapts the Company after each
|
|
stage. The world, people, tools and changing Company state are the primary
|
|
experience; UI is a compact field instrument attached to that world.
|
|
|
|
### Player Contract
|
|
|
|
The player repeatedly:
|
|
|
|
```text
|
|
read the current Company, Market and threat
|
|
-> choose a capability, Recruit, item, order or risk
|
|
-> commit one authoritative decision
|
|
-> observe automatic progress and its causal result
|
|
-> retain the Company growth and adapt the next stage
|
|
```
|
|
|
|
The first session teaches valuation, Company building, front-to-rear order,
|
|
risk and diagnosis in that order. A useful decision must change resources,
|
|
capability, legal options, risk, information or future Company shape; a new
|
|
row, currency or timer without a new decision is not valid scope.
|
|
|
|
### Persistent And Contract State
|
|
|
|
| Scope | Owns | On Contract completion/withdrawal |
|
|
| --- | --- | --- |
|
|
| Guild / Company | Recruit IDs, names, Profession, Trait, attributes, level, XP, equipment, history, Coin, Records, Renown and breadth | retained and available at Guild |
|
|
| Contract | Market offers, current Coin allocation, Line capacity, formation/risk snapshot, route, battle/result/recovery receipts and temporary effects | resolved or cleared through one receipt |
|
|
| Battle | current HP, event timeline, target state, pause/speed/inspect state and Knockout result | closed at Result/Recovery; no permanent Recruit deletion |
|
|
|
|
The initial four-offer choose-two decision runs once when New Game creates the
|
|
initial Company. A knocked-out Recruit leaves only the current Battle and
|
|
returns at the next stage with the same identity and persistent growth.
|
|
|
|
### Strategic Grammar
|
|
|
|
The long-term strategy is built from:
|
|
|
|
- Market valuation under Coin pressure;
|
|
- Profession, Trait and Equipment interactions;
|
|
- one-slot equipment tradeoffs and replacement inventory;
|
|
- Party Line order, protection, reach, healing and exposure;
|
|
- visible risk/reward and information confidence;
|
|
- deterministic automatic battle diagnosis;
|
|
- level/XP, promotion, roster breadth and later Contract choices.
|
|
|
|
Growth must keep opening new comparisons and combinations. It must not collapse
|
|
into a single dominant build or a sequence of unconditional numeric upgrades.
|
|
|
|
### Product Architecture
|
|
|
|
```text
|
|
Godot main / compatibility bridge
|
|
-> entry pages and input shell
|
|
-> domain services and authoritative receipts
|
|
-> content registry and deterministic offer generation
|
|
-> redundant local Save / migration / recovery
|
|
-> focused tests, normal-render evidence and human review
|
|
```
|
|
|
|
Pages render state and submit commands; services own rules and persistence;
|
|
content owns stable IDs and localized rows; tests observe the same player path.
|
|
No new page may implement its own Coin, Recruit, Contract or Save semantics.
|
|
|
|
### Authority Map
|
|
|
|
| Question | Authority |
|
|
| --- | --- |
|
|
| What the game is and why it continues | this document |
|
|
| Combat, Recruit, Profession, Party, failure and Save mechanics | `02_GAMEPLAY_AND_BALANCE.md` |
|
|
| Player pages, controls, states, localization and accessibility | `03_PAGES_AND_UX.md` |
|
|
| Japanese 2D visual identity, assets, animation and audio | `04_ART_ANIMATION_AUDIO.md` |
|
|
| Complete content identities and references | `05_CONTENT_CATALOG.md` |
|
|
| Ordered implementation, evidence and gates | `06_DEVELOPMENT_AND_REVIEW.md` |
|
|
|
|
When an older implementation conflicts with this positive contract, the
|
|
current Owner decision and these companion authorities win; the older behavior
|
|
is retained only as explicitly marked migration evidence.
|
|
|
|
<!-- Historical section source: GAME_PRODUCT_CONTRACT; retained in Git history. -->
|
|
|
|
## Aetherbound Guild: Game Product Contract
|
|
|
|
> Authority: player promise, complete loop, run and campaign shape, canonical
|
|
> vocabulary, product boundaries, and design acceptance.
|
|
>
|
|
> Revision: guild-charter-integration-r4 / 2026-08-11
|
|
>
|
|
> Companion authorities:
|
|
> [02_GAMEPLAY_AND_BALANCE.md](02_GAMEPLAY_AND_BALANCE.md),
|
|
> [02_GAMEPLAY_AND_BALANCE.md](02_GAMEPLAY_AND_BALANCE.md), and
|
|
> [02_GAMEPLAY_AND_BALANCE.md](02_GAMEPLAY_AND_BALANCE.md).
|
|
|
|
### 1. Authority And Decision Order
|
|
|
|
This document defines what the game promises. The systems authority owns
|
|
combat, party-line, recruit, profession, equipment, and artifact mechanics. The
|
|
economy authority owns prices, rewards, pacing, capacity selection, difficulty
|
|
numbers, and simulations. The save/failure authority owns commit, casualty,
|
|
interruption, retry, recovery, and rollback behavior.
|
|
|
|
When two statements conflict, apply this order:
|
|
|
|
1. the frozen repository specification and later explicit Owner decisions;
|
|
2. save integrity and exactly-once transaction rules;
|
|
3. formulas in the systems and economy authorities;
|
|
4. this product intent;
|
|
5. content, presentation, gallery, and prototype instances.
|
|
|
|
The repository remains design-only. No document in this package authorizes an
|
|
engine project, runtime implementation, production package, store operation,
|
|
or bulk final media production. Human comprehension, fun, visual, listening,
|
|
device, packaging, and release acceptance remain later gates.
|
|
|
|
#### 1.1 Owner-Locked Continuity And Knockout Rule
|
|
|
|
Owner decision, 2026-08-17, superseding the 2026-08-11 permanent-casualty
|
|
decision: **the Expedition Company is persistent across Contract runs and a
|
|
zero-HP Recruit is knocked out only for the current battle**. The Recruit leaves
|
|
the resolving line immediately and Reserve Recruits do not substitute during
|
|
that battle. At the next stage the knocked-out Recruit is battle-ready again
|
|
with the same stable ID, level, XP, Profession, Trait, equipment and history.
|
|
The exact stage-entry HP amount belongs to the later balance conversion; DEV-2
|
|
must not invent or persist a permanent casualty while that value is frozen.
|
|
|
|
Contract completion clears only Contract-local stage, route, formation snapshot
|
|
and temporary-effect state. It never discards the roster or its growth. Four
|
|
opening offers and the choose-two decision occur once when New Game establishes
|
|
the initial persistent Expedition Company.
|
|
|
|
### 2. Product Definition
|
|
|
|
**Aetherbound Guild** is a premium landscape 2D Japanese hand-drawn fantasy
|
|
incremental auto-battler about growing a persistent Expedition Company from
|
|
generated recruits. At each
|
|
market the player reads three battle offers, hires recruits, buys equipment,
|
|
locks or refreshes offers, equips the company, and orders one horizontal party
|
|
line from front to rear. The player commits to a risk, watches the deterministic
|
|
battle resolve, reads the causal result, then adapts after rewards, knockouts,
|
|
dismissals, recruitment and continuing growth.
|
|
|
|
The transferable play thesis is independently expressed as:
|
|
|
|
```text
|
|
read a market and a visible threat
|
|
-> buy a capability under price pressure
|
|
-> order one front-to-rear line
|
|
-> commit to a reward/risk contract
|
|
-> observe automatic execution
|
|
-> explain the first decisive interaction
|
|
-> keep, train, reserve, dismiss, or recruit
|
|
-> repeat with a changed company and budget
|
|
```
|
|
|
|
Player skill is preparation, valuation, formation order, risk selection, and
|
|
diagnosis. Battle controls are limited to start, pause, speed, inspect, and
|
|
retreat. There is no repeated manual skill activation and no mid-battle
|
|
formation or equipment editing.
|
|
|
|
#### 2.1 Audience
|
|
|
|
Primary players enjoy job systems, shop valuation, equipment interactions,
|
|
formation order, and watching a prepared plan work without attack-by-attack
|
|
input. Secondary players enjoy mastering a broad fixed catalog, finding hidden
|
|
hybrids, completing collections, and applying optional postgame laws.
|
|
|
|
The game must support 15-30 minute partial sessions, 45-75 minute Contract
|
|
expeditions inside one continuous save, touch-only play, frequent pausing, and
|
|
return after a long break.
|
|
|
|
#### 2.2 Commercial And Platform Boundary
|
|
|
|
| Field | Contract |
|
|
|---|---|
|
|
| Business model | One-time premium purchase; no ads, paid energy, random paid reward, paid currency, battle pass, daily streak, or pay-to-skip |
|
|
| Presentation | Original landscape 2D Japanese hand-drawn fantasy; no third-party commercial asset enters this repository |
|
|
| Target platforms | Landscape mobile, PC, and console with equivalent touch, mouse, keyboard, and controller semantics |
|
|
| Launch languages | English and Simplified Chinese from semantic message IDs, not concatenated strings |
|
|
| First clear | 15-25 authored hours across eight regions and sixteen contracts |
|
|
| Mastery | 60+ hours of profession, equipment, artifact, enemy, and high-difficulty breadth |
|
|
| Connectivity | Complete core play offline; platform services never gate play or save access |
|
|
| Development | Frozen until all design-completion gates, independent review, and Owner approval pass |
|
|
|
|
No real-time wait, offline income, social obligation, or monetized relief may be
|
|
introduced to solve economy or failure problems.
|
|
|
|
### 3. Canonical Vocabulary
|
|
|
|
These terms are shared keys. Product-facing documents must use them with the
|
|
same meaning.
|
|
|
|
| Term | Definition |
|
|
|---|---|
|
|
| Guild | Persistent save containing the Expedition Company, resources, unlock breadth, region progress, records, settings, and collection knowledge |
|
|
| Recruit | One persistent generated member of the Expedition Company, built from a generated name, modular appearance seed, fixed profession, rolled attributes, exactly one Trait, zero or one equipped item, an individual level/XP record, and battle history |
|
|
| Profession | One fixed reusable combat ruleset from the authored catalog |
|
|
| Trait | One short mechanical modifier with a visible benefit and cost; each Recruit has exactly one |
|
|
| Party line | The ordered simultaneous deployment from front to rear |
|
|
| Line capacity | Maximum allied footprint deployable at once for the current Contract expedition; distinct from persistent roster size |
|
|
| Reserve | A Recruit owned by the persistent Expedition Company but not currently deployed |
|
|
| Market | The between-battle shop state containing Recruit, equipment, and battle offers plus refresh (reroll) and lock actions |
|
|
| Battle offer | One combat offer with a fixed enemy order, risk tier, reward, retreat result, and seed |
|
|
| Initial Company draft | The durable pre-Guild record that reserves one New Game seed, its four opening Recruit offers, and the current accepted pair before the persistent Expedition Company exists |
|
|
| Round | One Market preparation, one committed battle, one result, and one reward/recovery handoff |
|
|
| Run | One bounded Contract expedition containing repeated Rounds and a Contract boss; it retains the persistent roster and growth when it ends |
|
|
| Contract | One twelve-Round regional challenge followed by a boss Muster and boss battle |
|
|
| Region | One campaign area containing two ordered Contracts and two bosses |
|
|
| Artifact | A run-defining shared rule modifier, distinct from unit equipment |
|
|
| Coin | Persistent earned operating currency unless a later named temporary resource explicitly says otherwise |
|
|
| Renown | Persistent breadth-unlock resource; it cannot buy raw combat power |
|
|
| Guild Charter | Persistent eight-Region board where Renown buys catalog, trial, threat-information, and optional Oath-law breadth for future generated runs |
|
|
| Knocked Out | A Recruit reduced to zero HP and removed from the current battle; the Recruit is battle-ready again for the next stage without losing persistent state |
|
|
| Meta progression | Persistent access, information, and rule breadth; it may not replace run decisions |
|
|
| Record | A persistent fact about discoveries, clears, mastery, or a past run; it never turns a Recruit into a fixed narrative lead |
|
|
|
|
`Expedition Company` describes all Recruits owned by the persistent Guild save.
|
|
It is not a deployment cap. Company size and Reserve count are independent from
|
|
the current Contract's Line capacity.
|
|
|
|
The canonical count used by every dismissal rule is:
|
|
|
|
```text
|
|
owned_recruit_count = count(unique Recruit IDs owned by the Expedition Company,
|
|
across deployed party and Reserve)
|
|
```
|
|
|
|
Offers, summons, dismissed Recruits and pending hires do not count. A knocked-
|
|
out Recruit remains owned and does count. A dismissal is legal only while a
|
|
Market is open and `owned_recruit_count >= 2` at both preview and commit. A
|
|
legal dismissal must leave at least one owned Recruit.
|
|
|
|
### 4. Fixed Content Envelope
|
|
|
|
The complete package contains exactly:
|
|
|
|
- 12 base Professions;
|
|
- 24 regular advanced Professions, exactly two descendants per base;
|
|
- 6 hidden hybrid Professions, for 42 total Profession identities;
|
|
- 36 mechanical Traits;
|
|
- 320 equipment items;
|
|
- 60 Artifacts;
|
|
- 100 ordinary/elite enemy identities plus 16 bosses;
|
|
- 8 Regions, each with two Contracts and two bosses.
|
|
|
|
Recruits are generated compositional units, not an authored cast. They do not
|
|
need recruit-specific social progression, courtship, individual narrative
|
|
arcs, or individual voice-content systems. Generated names and modular visual
|
|
parts are presentation combinations, not additional gameplay identities.
|
|
|
|
Every catalog row must alter a decision. A row that differs only by name,
|
|
rarity, color, art, or a scalar is filler and fails the design gate.
|
|
|
|
### 5. Player Promise And Experience Pillars
|
|
|
|
#### 5.1 Order Is A Build Axis
|
|
|
|
Front-to-rear order changes legal targets, protection, reach, projectile
|
|
collision, healing access, displacement, summon space, and line collapse after
|
|
a knockout. Reordering one pair must visibly change at least one predicted event in
|
|
a valid fixture.
|
|
|
|
#### 5.2 Automatic Battle Is Trustworthy
|
|
|
|
The same initial state, content revision, seed, and retreat timestamp produces
|
|
the same result. Major actions, target changes, immunity, displacement, and
|
|
failed actions have inspectable reasons. Presentation speed never changes the
|
|
simulation.
|
|
|
|
#### 5.3 Defeat Interrupts Progress Without Deleting Growth
|
|
|
|
A knockout removes the Recruit's contribution for the rest of the current
|
|
battle and may reduce reward, mastery rating or route momentum, but it never
|
|
deletes the Recruit or equipment. The next stage makes that Recruit battle-ready
|
|
again under the later balance authority's stage-entry HP rule. Dismissal remains
|
|
the only ordinary roster-removal operation and cannot be used for profit.
|
|
|
|
#### 5.4 Randomness Presents Bounded Markets
|
|
|
|
Seeded offers vary a run, but every Market contains a steady battle option and
|
|
the recovery rules preserve an affordable base Recruit after ordinary losses.
|
|
Locks, escalating refresh prices, fixed catalog schemas, and separated random
|
|
streams prevent closing, reloading, or unrelated draws from changing an offer.
|
|
|
|
#### 5.5 Progression Adds Grammar
|
|
|
|
Across Contract runs, persistent levels, Profession promotions, equipment,
|
|
Artifacts and recruited members expand planning. Guild progression also opens
|
|
new sidegrades, previews, challenges and catalog breadth. Growth must keep
|
|
revealing new decisions rather than becoming an unconditional stat-only ladder.
|
|
|
|
### 6. Product Vetoes
|
|
|
|
Reject a feature or content row if it depends on:
|
|
|
|
- manual attack or skill spam during battle;
|
|
- a second battlefield axis, hidden sub-row, or free actor wandering;
|
|
- silent targeting, reach, protection, immunity, or retarget exceptions;
|
|
- a universally correct item, Trait, Profession, order, hire, or refresh plan;
|
|
- a generated Recruit needing authored individual plot or bespoke voice scope;
|
|
- permanent power grinding to make Standard difficulty fair;
|
|
- paid rescue, paid reroll, real-time recovery, or offline resource gain;
|
|
- deletion without an explicit transaction and causal record;
|
|
- a battle over 150 simulation seconds;
|
|
- content volume, rarity, or cosmetic variants used to conceal filler;
|
|
- copied names, prose, formulas, layouts, trees, art, audio, or source
|
|
expression from private reference material.
|
|
|
|
### 7. Complete Loop And State Handoffs
|
|
|
|
#### 7.1 New Game To Persistent Company, Then Contract Run
|
|
|
|
```text
|
|
durably reserve one New Game seed in an Initial Company draft
|
|
-> generate four provisional Recruits from that seed at offer indices 0..3
|
|
-> choose exactly two at no Coin cost, revising the pair if desired
|
|
-> atomically commit those two Recruit records into the persistent Expedition Company
|
|
-> enter the Guild
|
|
-> choose an unlocked Contract and create its bounded expedition state
|
|
```
|
|
|
|
Seed reservation and offer materialization are persistence steps inside one
|
|
player-facing Initial Company page. Once the four offers exist, Back, Cancel,
|
|
Save and Quit, process restart, clock change or repeated confirm returns the
|
|
same draft and cannot allocate another seed. The player may revise the accepted
|
|
pair until Company commit. That commit promotes the draft exactly once; only
|
|
then are the two unchosen provisional offers discarded. The chosen two become
|
|
persistent Recruits. The first Market allows at least two affordable hires so
|
|
the player can expand the opening roster without preselecting a build.
|
|
|
|
#### 7.2 Repeated Round
|
|
|
|
```text
|
|
open the stable Market
|
|
-> inspect three Recruit, four equipment, and three battle offers
|
|
-> buy, sell, dismiss, lock, or refresh
|
|
-> equip and order the Party line
|
|
-> choose one battle offer and review exact failure exposure
|
|
-> commit and resolve automatic battle
|
|
-> inspect first decisive cause and all knocked-out Recruits
|
|
-> claim one fixed reward or resolve defeat/retreat
|
|
-> increase capacity at the declared clear milestones
|
|
-> make knocked-out Recruits battle-ready for the next stage
|
|
-> open the next Market with the persistent Expedition Company and earned resources
|
|
```
|
|
|
|
The player may hire beyond current Line capacity. Those Recruits enter Reserve
|
|
and remain available for later preparation. No Reserve enters the current
|
|
battle automatically after a knockout.
|
|
|
|
#### 7.3 Boss Muster And Contract Return
|
|
|
|
After twelve cleared Rounds, capacity reaches ten and the Boss Muster opens a
|
|
final non-rerollable market. The boss preview exposes its target, phase,
|
|
summon, displacement, and timeout rules. Boss victory clears the Contract.
|
|
|
|
A Contract run ends by clear or explicit withdrawal. The result commits Region
|
|
progress, rewards, records and the next unlocked decision, then returns the
|
|
persistent Expedition Company to the Guild. Recruits, levels, XP, Professions,
|
|
Traits, equipment and earned persistent resources remain playable. Only
|
|
Contract-local stage, route, battle snapshot and temporary-effect state closes.
|
|
|
|
The sixteenth campaign boss uses the same run-close settlement, then opens one
|
|
explicit first-clear decision only after every result, casualty, reward, and
|
|
progression receipt is verified:
|
|
|
|
- **Complete the First Chronicle** records the clear, presents the finite ending
|
|
and Final Ledger, then returns to the persistent Guild/Title boundary.
|
|
- **Open the Oath Season** keeps that same Expedition Company and opens the
|
|
postgame hub under separately reviewed optional laws.
|
|
|
|
Both choices settle the cleared Contract state exactly once. Neither choice
|
|
duplicates or deletes a Recruit, equipment item, level, XP or earned persistent
|
|
resource.
|
|
|
|
### 8. Recruitment, Dismissal, Knockout, And Continuity Promise
|
|
|
|
- Each Recruit has one current Profession. It cannot freely change jobs.
|
|
- A base Recruit may make one persistent promotion when the
|
|
requirements in the systems authority are met.
|
|
- All rolled attribute totals at a given offer level use the same budget;
|
|
distribution, Profession, Trait, order, and price create value differences.
|
|
- A Recruit has exactly one Trait for its entire lifetime.
|
|
- A Recruit has exactly one universal equipment slot. Item form is eligibility
|
|
and presentation, not an additional simultaneous slot; equipping a candidate
|
|
atomically returns the previous item, if any, to persistent Guild inventory.
|
|
- Level and XP belong to one Recruit, are bounded to levels `1..10`, and never
|
|
become a shared Company level or transfer to another Recruit.
|
|
- Dismissal is legal only in a Market when the canonical
|
|
`owned_recruit_count >= 2`, returns equipment, removes the Recruit, and
|
|
requires review when survival history is nonzero. Preview and commit both
|
|
revalidate the count.
|
|
- Knockout never selects, destroys or unequips the equipped item.
|
|
- Knockout grants no same-battle respawn, resurrection timer, extra life or
|
|
automatic Reserve substitution.
|
|
- At the next stage the Recruit is battle-ready again with identical persistent
|
|
identity and growth state; the later balance conversion owns the exact HP
|
|
restored.
|
|
- All hire, dismissal, knockout outcome and recovery changes are atomic
|
|
operations owned by the save/failure authority.
|
|
|
|
Attempting to dismiss the final owned Recruit is blocked with
|
|
`FINAL_OWNED_RECRUIT`. It removes nothing and returns no rebate. Contract
|
|
withdrawal is a separate expedition operation and never deletes the persistent
|
|
roster.
|
|
|
|
This is a persistent generated-roster growth model, not a fixed-protagonist or
|
|
character-bond simulator. Result reports respect generated identity and battle
|
|
history without creating authored biography obligations.
|
|
|
|
### 9. Capacity And Cumulative Recruitment
|
|
|
|
Every Contract expedition begins at four footprint and increases after cleared-
|
|
Round counts `2, 4, 6, 8, 10, 12`, reaching ten for the boss. The formula and
|
|
five-profile selection evidence are solely in the economy authority.
|
|
|
|
Line capacity limits simultaneous allied footprint, not purchases. The Guild
|
|
and active Contract store separate values for:
|
|
|
|
```text
|
|
line_capacity
|
|
owned_recruit_ids[]
|
|
deployed_recruit_ids[]
|
|
reserve_recruit_ids[]
|
|
recruits_hired_total
|
|
recruit_knockout_events_total
|
|
recruits_dismissed_total
|
|
```
|
|
|
|
There is no cumulative hire cap. Market supply, Coin, escalating refresh cost,
|
|
and low dismissal recovery are the anti-hoarding constraints. A content row or
|
|
UI label may not treat ten as the run roster total.
|
|
|
|
### 10. Campaign, Session, And Decision Shape
|
|
|
|
| Scope | Target | Terminal state |
|
|
|---|---:|---|
|
|
| Battle | 25-110 s typical; 150 s hard cap | Causal report, reward, defeat, or retreat |
|
|
| Round | 3-6 min | Stable next Market with committed result, recovery and reward |
|
|
| Contract run | 45-75 min; twelve Rounds plus boss | Clear or explicit withdrawal, then Guild return |
|
|
| Typical session | 15-75 min | Suspend-safe Market, battle snapshot, or run result |
|
|
| First clear | 15-25 h; sixteen ordered Contracts across eight Regions | Finite ending and postgame selection |
|
|
| Mastery | 60+ h | Visible Profession, catalog, boss, and challenge completion records |
|
|
|
|
Each Region's first Contract teaches one new pressure and its second combines
|
|
that pressure with earlier rules. Each of the 16 bosses synthesizes the current
|
|
Contract rather than merely increasing HP. Unlock order must provide at least
|
|
two rational counters to every mandatory boss without requiring a random item.
|
|
|
|
The five pacing profiles, exact Market/battle/decision timestamps, full-clear
|
|
budgets, and no-decision gaps are owned by the economy authority. A battle can
|
|
be fully automatic while remaining decision-dense because a causal result and
|
|
the next stable Market appear within 15 seconds of resolution.
|
|
|
|
### 11. Onboarding Contract
|
|
|
|
Teaching uses **safe action -> guided comparison -> independent commit**. It
|
|
must never make the entire opening build for the player.
|
|
|
|
| Target | Required player action | Safe boundary | Understanding proof |
|
|
|---:|---|---|---|
|
|
| 0:00-1:00 | Choose two provisional Recruits from four generated offers | Pair can be revised before atomic Company commit; return/restart restores the same four | States one Profession job and one Trait cost |
|
|
| 1:00-3:00 | Hire two Recruits and fill capacity four | First Market has protected affordable offers | Distinguishes Line capacity from owned Company |
|
|
| 3:00-5:00 | Swap front and rear positions | Predicted targets update before commit | Identifies who protects and who can reach |
|
|
| 5:00-7:00 | Choose steady or pressured battle offer | Both are completable; risks are explicit | States reward difference and first threat |
|
|
| 7:00-9:00 | Start and inspect automatic battle | Pause and 1x remain available | Identifies first decisive interaction |
|
|
| 9:00-12:00 | Choose reward and adapt next Market | Reward draft is fixed; no reload reroll | Makes one causal equipment/order change |
|
|
| 12:00-16:00 | Recover after a scripted nonterminal knockout fixture | Training state grants no persistent reward or roster loss | Explains current-battle removal and next-stage return |
|
|
| 16:00-20:00 | Complete an independent Round | One no-reward reset before commit | Buys, orders, commits, and diagnoses without a pointer |
|
|
|
|
Experienced players may skip highlights, but the concise rule reference remains
|
|
available. Returning players receive a state recap, next objective, current
|
|
Market/encounter boundary, knockouts, and one optional no-reward rehearsal.
|
|
|
|
### 12. Difficulty And Postgame Promise
|
|
|
|
Difficulty changes enemy and reward parameters, not hidden target rules,
|
|
catalog access, save integrity, or persistent Recruit ownership. Standard is
|
|
balanced without Meta power. Lower difficulty preserves all systems and
|
|
Records. Higher difficulty rewards broader risk selection but cannot become a
|
|
mandatory progression route.
|
|
|
|
Persistent breadth is reviewed through the **Guild Charter**, an original board
|
|
of eight horizontal Region seals and prerequisite tracks. It spends Renown only
|
|
on catalog packets, Profession clues or trials, threat information, and optional
|
|
Oath laws. It never contains percentage power, starting resources, extra Line
|
|
capacity, Recruit preservation, or casualty reversal.
|
|
|
|
The sixteenth boss completes the campaign before asking the player to choose:
|
|
|
|
- **Complete the First Chronicle:** record the clear, view the finite ending and
|
|
Final Ledger, and return to the persistent Guild/Title boundary.
|
|
- **Open the Oath Season:** keep the same Expedition Company, enter the postgame
|
|
hub, and review optional laws for further Contract expeditions.
|
|
|
|
After that finite ending, postgame provides:
|
|
|
|
- **Profession Mastery:** clear fixed counterfactual trials for every base,
|
|
regular advanced, and hidden hybrid identity;
|
|
- **Catalog Completion:** discover and validate every equipment, Artifact,
|
|
Trait, ordinary/elite enemy, and boss identity;
|
|
- **Oath Contracts:** apply one or more fully previewed run laws that change
|
|
valuation, formation pressure, market constraints, or enemy composition;
|
|
- **Final Ledger:** a finite completion screen for all mastery and challenge
|
|
requirements.
|
|
|
|
Postgame does not promise endless content or infinite stat scaling. Repeated
|
|
unchanged clears cannot be the majority of the 60+ hour mastery target. The
|
|
cleared Expedition Company remains the playable persistent roster; Records
|
|
reference its history without cloning or re-instantiating Recruit IDs.
|
|
|
|
### 13. Accessibility, Localization, And Input
|
|
|
|
#### 13.1 Accessibility Baseline
|
|
|
|
- UI scale presets at 100%, 115%, and 130% without clipped functional text;
|
|
- reference landscape layouts at small, standard, and large safe areas;
|
|
- reduced motion, reduced flash, camera-shake intensity, and hit-stop options;
|
|
- independent master, music, ambience, voice, and effects volume;
|
|
- captions for every information-bearing audio event;
|
|
- color-independent rarity, target, protection, status, risk, and casualty
|
|
encoding;
|
|
- high-contrast target paths and Party-line position indicators;
|
|
- pause, 1x, 2x, and 4x without loss of event-log information;
|
|
- complete keyboard/controller remapping and complete touch path;
|
|
- screen-reader labels and deterministic focus order for every Market,
|
|
formation, result, recovery, and confirmation state;
|
|
- no required rapid tap, precision drag, simultaneous input, audio-only cue, or
|
|
unpausable reading window;
|
|
- low-power presentation settings that never alter simulation ticks or results.
|
|
|
|
#### 13.2 Input Semantics
|
|
|
|
All platforms map the same verbs: focus, inspect, compare, buy, sell, lock,
|
|
refresh, equip, reorder, confirm, cancel, start, pause, speed, retreat, and open
|
|
event log. Touch instructions name Tap as the primary operation: Tap a Recruit
|
|
or item to select, then Tap a legal destination or target to commit the pending
|
|
choice. Drag is an optional shortcut; select-then-destination performs every
|
|
reorder and equip action. Battle provides no hidden pointer-only command
|
|
surface.
|
|
|
|
Destructive or spending actions use a review state. Duplicate confirm input
|
|
returns the first operation receipt and cannot spend, dismiss, claim, or close
|
|
a run twice.
|
|
|
|
#### 13.3 Localization Rules
|
|
|
|
- English and Simplified Chinese share semantic message IDs and typed values.
|
|
- No sentence is assembled from translated fragments.
|
|
- Generated names store component IDs and a rendered snapshot so migration and
|
|
language changes remain deterministic.
|
|
- System nouns have one glossary entry; content cannot invent synonyms for the
|
|
canonical state terms.
|
|
- Layout validation covers English pseudolocalization at 130% expansion and
|
|
Simplified Chinese at 130% UI scale.
|
|
- Player-entered profile/run labels are Unicode-safe; generated Recruit names
|
|
remain content-filtered and do not require online moderation.
|
|
|
|
### 14. Content Interface And No-Filler Gate
|
|
|
|
Every consumed gameplay row declares:
|
|
|
|
```text
|
|
stable_id and localized_name_id
|
|
catalog family and unlock placement
|
|
decision changed
|
|
state inputs read and authoritative outputs changed
|
|
trigger, timing, target, tradeoff, and counterplay
|
|
at least two nontrivial interactions
|
|
failure mode and causal-report event
|
|
presentation, accessibility, localization, and asset needs
|
|
fixed validation fixtures and expected event differences
|
|
```
|
|
|
|
A Recruit offer additionally declares Profession, attribute budget/distribution,
|
|
one Trait, individual level/XP, one starting item, generated identity seed,
|
|
list price, and offer provenance. An
|
|
opening provisional offer also declares its Initial Company draft ID, fixed index
|
|
`0..3`, deterministic offer ID, and canonical output hash. A
|
|
Profession changes battlefield responsibility or resource/timing/target/order
|
|
grammar. Equipment and Artifacts change behavior, not only magnitude. An enemy
|
|
creates a distinct preview-to-counter question. A Region changes composition,
|
|
market pressure, or formation valuation. Any row failing its decision test is
|
|
rejected before production planning.
|
|
|
|
### 15. Product Validation Gates
|
|
|
|
#### 15.1 Comprehension
|
|
|
|
With at least five uncoached target players:
|
|
|
|
- four identify front and rear and predict the first legal target;
|
|
- four distinguish Line capacity, persistent Company, and cumulative hires;
|
|
- four explain one equipment or order counterfactual before battle;
|
|
- four state what knockout, dismissal, retreat, and Contract return preserve;
|
|
- four can explain why the final owned Recruit cannot be dismissed and can
|
|
find the separate `Withdraw Contract` action;
|
|
- four use refresh and lock without expecting a reload reroll;
|
|
- four can state what Renown buys on the Guild Charter and why no claim adds
|
|
permanent combat power;
|
|
- four distinguish Complete the First Chronicle from Open the Oath Season and
|
|
predict that both preserve the Expedition Company while changing destination;
|
|
- three voluntarily start another Round or name a specific build test.
|
|
|
|
#### 15.2 Decision And Economy Quality
|
|
|
|
- A meaningful hire or provisional-Recruit choice occurs within one minute.
|
|
- A formation-order counterfactual occurs within five minutes.
|
|
- Every 20-minute profile trace stays within its stated decision-gap bound.
|
|
- Capacity ten passes all five analytical profile gates; larger candidates do
|
|
not pass the same complete gate set.
|
|
- Three ordinary knockouts return the same Recruit IDs for the next stage in the
|
|
fixed recovery fixture without duplicating value.
|
|
- Fixed seeds reject always-refresh, always-hire and always-dismiss as dominant
|
|
strategies.
|
|
|
|
#### 15.3 Battle Trust
|
|
|
|
- Reorder fixtures change targeting, reach, protection, healing, movement, or
|
|
projectile results exactly as previewed.
|
|
- Every rejected action and retarget has one machine-readable reason.
|
|
- 1x exposes every major cast, projectile collision, displacement, summon,
|
|
knockout, line-collapse, and retreat event.
|
|
- 4x and pause produce the same outcome hash as uninterrupted 1x.
|
|
- The result identifies the first decisive cause before aggregate statistics.
|
|
|
|
#### 15.4 Open Human Gates
|
|
|
|
Machine checks cannot approve fun, emotional weight, visual clarity, listening
|
|
quality, generated-name quality, device ergonomics, or willingness to continue.
|
|
Those require prototype, target-player, device, independent-review, and Owner
|
|
evidence. Runtime, media production, packaging, store, and release remain
|
|
separately authorized.
|
|
|
|
### 16. Acceptance Checklist
|
|
|
|
- [ ] One ordered Party line is the only deployment grammar.
|
|
- [ ] Battle is fully automatic apart from diagnostic controls and retreat.
|
|
- [ ] Generated-Recruit, knockout, next-stage return, dismissal, Reserve,
|
|
capacity, and cumulative-hire rules are explicit.
|
|
- [ ] Final-owned-Recruit dismissal is illegal at preview and commit without
|
|
changing Contract withdrawal or knockout recovery.
|
|
- [ ] All four authorities use the same counts, terms, IDs, and ownership.
|
|
- [ ] The complete Market loop includes Recruit, equipment, battle, refresh,
|
|
lock, risk, result, reward, and recovery decisions.
|
|
- [ ] Initial Company creation reserves its seed before generating the four
|
|
opening offers, preserves a revisable exact pair, and promotes the draft
|
|
into the persistent roster exactly once.
|
|
- [ ] First clear is 15-25 hours and mastery is 60+ hours without waiting.
|
|
- [ ] The Guild Charter has eight horizontal Region seals, sourced
|
|
prerequisites, existing Renown costs, and breadth-only effects.
|
|
- [ ] The final boss presents Chronicle completion and Oath Season as two
|
|
explicit choices; both preserve the Expedition Company without duplicating
|
|
Recruit IDs.
|
|
- [ ] Every gameplay row passes the no-filler decision test.
|
|
- [ ] Accessibility, localization, offline, premium, and save-integrity
|
|
boundaries are preserved.
|
|
- [ ] Originality, human acceptance, development, and release gates remain open.
|
|
|
|
<!-- Historical section source: ORIGINALITY_BOUNDARY; retained in Git history. -->
|
|
|
|
## Aetherbound Guild Originality Boundary
|
|
|
|
### Original Product
|
|
|
|
Aetherbound Guild may study genre structures such as automatic party combat,
|
|
shop valuation, one-dimensional formation, equipment tradeoffs and visible
|
|
enemy telegraphs. Its world, names, prose, formulas, progression topology,
|
|
screen compositions, characters, enemies, art, animation, music, sound and code
|
|
must be independently authored.
|
|
|
|
### Prohibited Reproduction
|
|
|
|
- No commercial game code, extracted asset, music, sound, font or resource pack
|
|
enters the product repository.
|
|
- No reference screenshot becomes a traced or near-identical product layout.
|
|
- No reference name, description, formula, profession tree, enemy identity,
|
|
level sequence or reward table is copied.
|
|
- No downloaded commercial material is used as an image-generation reference.
|
|
|
|
### Product-Specific Decisions
|
|
|
|
- Recruits are generated, recruitable and dismissible persistent members; they
|
|
have no fixed lead cast or authored personal narrative system.
|
|
- Battle uses one continuous front-to-rear line, not a three-lane grid.
|
|
- The player prepares and diagnoses automatic battle; repeated manual attack or
|
|
skill activation is not part of the product.
|
|
- The visual identity is original Japanese 2D hand-drawn fantasy, not copied
|
|
commercial art and not the rejected pixel implementation.
|
|
|
|
### Evidence
|
|
|
|
Reference research and rejected attempts may remain in Git history or private
|
|
research storage. They are not current product authority. Runtime production
|
|
records retain only the provenance needed to prove that accepted assets and
|
|
audio are owned, licensed or independently generated.
|