35 KiB
35 KiB
Authorized Reference Runtime Census
Purpose And Boundary
This document records the strongest content census supported by the authorized private archive. It is a research record for Aetherbound Guild, not a reference game encyclopedia. It deliberately omits commercial names, prose, formulas, layouts, art, audio and source expression.
The census distinguishes player-visible gameplay identities from files, components, animation frames and presentation variants. No extracted commercial asset is stored in this repository. All evidence links point to the authorized local archive.
Counting Rules
- Direct observation means a count made from captured runtime UI or from a deterministic package inventory. The method column says which.
- Inference means the evidence supports a structural conclusion but does not expose a player-facing record or total.
- Unknown means no defensible numerical total is available.
- Exact always applies only to the stated observation scope. An exact count on one captured screen or one inspected build is not automatically an exact game-wide total.
- Lower bound means the archive proves at least that many distinct entries; unseen content may exist.
- High confidence means the evidence can be re-counted directly. Medium confidence means classification or de-duplication requires judgment. Low confidence is not used as a substitute for an unsupported count; such a total is reported as unknown.
- Names that recur at different rarities, levels or capture times are counted once unless the visible behavior establishes a separate identity.
Provenance Index
The archive-level sources are:
- P-OBS: ordinary-play scope, player loop and unresolved runtime surfaces.
- P-AUDIT: the prior structural lower-bound audit and database-shell finding.
- P-EXT: deterministic Unity export totals and the clean-room boundary.
- P-IMG: the complete image catalog and asset-category assignments.
- P-CODE: type inventory and IL2CPP/source-code boundary.
- P-MANIFEST: extraction manifest for build and export traceability.
- P-VID: the 2026-08-11 fixed user-play video study and its private evidence package.
- P-PH3: the 2026-08-11 public-video Phase 3 reconstruction, with three independently bound evidence lanes and an explicit version boundary.
- P-PH4: the targeted replication-readiness closure covering Rune spending, mobile tutorial input, boss escalation, final-boss choice and retirement UI.
The screenshot groups used for manual visible-identity counts are:
- Recruits and profession presentations: P-RCR-01, P-RCR-02, P-RCR-03, P-RCR-04, P-RCR-05, P-RCR-06, P-RCR-07 and P-RCR-08.
- Skill tree: P-SKL-00, P-SKL-01, P-SKL-02, P-SKL-03, P-SKL-04, P-SKL-05, P-SKL-06, P-SKL-07 and P-SKL-08.
- Enemy roles: P-ENM-01, P-ENM-02, P-ENM-03, P-ENM-04, P-ENM-05 and P-ENM-06.
- Equipment and item presentations: P-ITM-01, P-ITM-02, P-ITM-03, P-ITM-04, P-ITM-05, P-ITM-06, P-ITM-07, P-ITM-08, P-ITM-09, P-ITM-10, P-ITM-11, P-ITM-12, P-ITM-13, P-ITM-14, P-ITM-15, P-ITM-16, P-ITM-17, P-ITM-18 and P-ITM-19.
- Milestone artifacts: P-ART-01, P-ART-02, P-ART-03 and P-ART-04.
- Formation and party equipment: P-FRM-01 and P-FRM-02.
- Run scope and mode surface: P-RUN-01, P-RUN-02 and P-RUN-03.
Screenshot filenames are provenance locators only. Their embedded ordinal text must not be interpreted as a game-wide content total.
Runtime And Content Census
Each row supplies the required evidence class, exactness, confidence, observation method, provenance and unresolved boundary.
| ID | Content surface | Count or result | Evidence class and exactness | Confidence | Observation method | Provenance | Why the game-wide total remains unresolved |
|---|---|---|---|---|---|---|---|
| C-01 | Research play records | 2 investigator-captured sessions, 1 fixed user-play recording and 8 public-video visual sources | Direct runtime observation; exact for current archive scope, not unique game builds | High | Kept each source, clip range, version boundary and terminal separate | P-OBS, P-VID, P-PH3, P-PH4 | The archive is not an exhaustive survey of all saves, modes or versions; two additional public sources supplied metadata/captions only and no visual mechanic claim. |
| C-02 | Named recruit labels | At least 8 distinct visible labels | Direct runtime observation; lower bound | High | Manually transcribed and de-duplicated visible recruit and party labels across both captured sessions; ignored repeated appearances and equipment variants | P-RCR-01 through P-RCR-08 | No complete roster or collection surface was captured; procedural label generation also cannot be excluded. |
| C-03 | Base profession identities | At least 7 identities | Direct runtime observation; lower bound | High | Re-counted the distinct base-role identities consolidated from ordinary-play tooltips and the visible progression surface | P-AUDIT, P-RCR-01 through P-RCR-08, P-SKL-00 | No complete profession index was captured, and the package database does not expose its runtime array. |
| C-04 | Advanced profession identities | At least 6 identities | Direct runtime observation; lower bound | High | Re-counted distinct advanced-role labels exposed in the captured progression and party evidence | P-AUDIT, P-SKL-00 through P-SKL-06 | Unseen, locked or build-gated professions cannot be counted from the captured surface. |
| C-05 | Progression branches | 6 visible branches | Direct runtime observation; exact for the captured tree surface and a lower bound for the build | High | Counted the persistent tree's visibly separated branch columns and opened one representative tooltip per branch | P-SKL-00 through P-SKL-06, P-OBS | Hidden, disabled or later-version branches cannot be ruled out. |
| C-06 | Visible progression slots | 28 slot/icon identities | Direct runtime observation corroborated by direct package inventory; exact for the captured tree namespace, not an active-skill total | High | Counted visible tree positions and verified a contiguous catalog namespace from skill_00 through skill_27; paired texture and sprite exports were de-duplicated by name |
P-SKL-00 through P-SKL-08, P-IMG | A slot can be locked, passive, repeated, decorative or unavailable; no serialized skill-record array was exposed. |
| C-07 | Active skill definitions | Unknown | Unknown; no exact or lower-bound entity total asserted beyond C-06's visible slots | High confidence in the unknown status | Looked for a player-facing index and a serialized skill array, then checked the type and image inventories | P-SKL-00, P-AUDIT, P-CODE, P-IMG | Slot art and type identities do not expose enabled runtime definitions, ranks or build-gated content. |
| C-08 | Inspectable ordinary enemy roles | At least 4 role presentations | Direct runtime observation; lower bound | High | Compared inspectable battlefield tooltips and silhouettes and de-duplicated them by combat role: ranged pressure, shielded frontline, reach frontline and caster pressure | P-ENM-01 through P-ENM-06 | The captures cover only early ordinary encounters and do not expose an enemy collection. |
| C-09 | Enemy combat identities | Unknown | Unknown; C-08 is a role lower bound, not an identity total | High confidence in the unknown status | Cross-checked runtime captures, enemy-category images and the empty serialized database surface | P-ENM-01 through P-ENM-06, P-AUDIT, P-IMG | Layered bodies, faces, weapons and effects cannot be mapped one-to-one to enemy records; elite and boss totals were not enumerated. |
| C-10 | Equipment and item labels | At least 16 distinct player-facing labels | Direct runtime observation; lower bound | High | Manually de-duplicated tooltip labels across opening, tutorial, shop, loot and backpack captures; repeated items, rarity changes and equipped views were counted once | P-ITM-01 through P-ITM-19 | No complete item collection was captured, and the runtime item arrays are not serialized in the extracted shells. |
| C-11 | Equipment and item records | Unknown | Unknown; no exact game-wide total | High confidence in the unknown status | Compared the visible lower bound with the equipment image category and database components | P-AUDIT, P-IMG, P-CODE | An item can use multiple layers or rarity presentations, while several records may share art; neither direction supports a one-to-one count. |
| C-12 | Milestone artifact choices | 3 distinct choices on one captured milestone | Direct runtime observation; exact for that choice set and a lower bound for the build | High | Opened each visible option tooltip, de-duplicated by conditional behavior, then verified the selected secondary-slot state | P-ART-01 through P-ART-04, P-OBS | Only one milestone choice was captured; the full artifact collection and duplicate rules were not enumerated. |
| C-13 | Artifact records | Unknown | Unknown; no exact game-wide total | High confidence in the unknown status | Compared the captured choice with the artifact-system type and its non-enumerable component shell | P-ART-01 through P-ART-04, P-AUDIT, P-CODE | The package proves the system exists but exposes no custom serialized artifact array. |
| C-14 | Editable formation grammar | At least 1 ordered party-line formation surface | Direct runtime observation; lower bound | High | Observed party cards being rearranged before combat and the resulting battlefield order | P-FRM-01, P-FRM-02, P-OBS | No formation index, preset list or complete capacity progression was captured. Arbitrary arrangements are states, not necessarily authored formation identities. |
| C-15 | Authored formations or presets | Unknown | Unknown; no numerical total | High confidence in the unknown status | Searched captured screens and structural inventory for an enumerable formation catalog | P-FRM-01, P-OBS, P-CODE | The evidence proves rearrangement, not a finite catalog of named formations. |
| C-16 | Environment or biome families | At least 5 visual asset families | Inference from direct package inspection; lower bound | Medium | Grouped environment exports by recurring visual family and checked the prior structural audit | P-AUDIT, P-IMG, P-EXT | Asset families may be unused, transitional, layered or combined at runtime; later traversal was not captured. This is not a runtime biome count. |
| C-17 | Victories at observed run terminals | 12 in an earlier completed adventure; separate party-wipe terminals at 30, 59 and 126 accumulated victories | Direct runtime observation; each value exact only for its recorded run | High | Read each fixed terminal state without combining counters across runs, modes or versions | P-OBS, P-RUN-01, P-RUN-02, P-VID, P-PH3 | Victory count does not reveal chapter count, wave count, encounter taxonomy, completion length or a game-wide cap. |
| C-18 | Chapters | Unknown | Unknown; no exact or lower-bound chapter total asserted | High confidence in the unknown status | Checked the captured run, environment catalog and type inventory for a chapter index | P-OBS, P-AUDIT, P-IMG, P-CODE | No chapter-select or complete campaign surface was captured, and environment families are not chapters. |
| C-19 | Waves or encounters | Unknown | Unknown; no exact game-wide total | High confidence in the unknown status | Reviewed the completed-run ledger and combat captures without treating screenshot ordinals or victories as authored wave totals | P-OBS, P-RUN-01, P-RUN-02 | Encounters can contain multiple waves, optional branches or special events; the archive does not expose the mapping. |
| C-20 | Achievements | Unknown | Unknown; no numerical total | High confidence in the unknown status | Searched captured surface inventory and type inventory for an enumerable achievement page or record set | P-OBS, P-CODE | No achievement surface or serialized achievement catalog was captured. |
| C-21 | Collection entries | Unknown | Unknown; no numerical total | High confidence in the unknown status | Searched captured surface inventory for roster, enemy, item, artifact or general collection indexes | P-OBS, P-AUDIT | No complete collection surface was captured; separate visible samples cannot establish completion totals. |
| C-22 | Play modes | At least 1 run-based adventure path | Direct runtime observation; lower bound | High | Inspected the new-run selection and setup states and separated the optional tutorial from the adventure path | P-RUN-03, P-OBS | The screen does not establish every unlockable or later-version mode. Battle speed and zoom are controls, not modes, so they are excluded. |
| C-23 | Exact mode total | Unknown | Unknown; no exact game-wide total | High confidence in the unknown status | Cross-checked the visible run selector with the captured surface inventory | P-RUN-03, P-OBS | Locked, postgame, challenge or build-gated paths were not enumerated. |
| C-24 | Displayed roster and field scale | Capacity states through 10/10; ten members simultaneously fielded in one late run |
Direct runtime observation; exact for the named frames and a lower bound for field scale | High | Reconciled early 7/10 and 8/10 states with a later 10/10 roster and ten live member panels |
P-VID, P-PH3 | The evidence does not prove that 10 is universal across every mode/version or a total recruit-identity count. |
| C-25 | Inventory and storage scale | Carried capacity 22 plus a separate storage surface observed at 1/10 and 7/10 |
Direct runtime observation; exact for the displayed states | High | Compared carried-inventory headers with the distinct storage grids in separate source-bound runs | P-VID, P-PH3 | Capacity upgrades, all contents, transfer input, persistence and cross-mode limits remain unknown. |
| C-26 | Shop progression surface | Visible tiers through 10; ordinary three-offer and broader states with at least eight visible offers | Direct runtime observation; lower bound for shop-state breadth | High | Read fixed headers and separately classified recurring offer layouts and upgrade-information states | P-VID, P-PH3 | Upgrade formula, complete tier range, quality probabilities, refresh rules and full offer pools remain unknown. |
| C-27 | Runtime arena presentations | At least 4 directly seen families: forest, cave, red-rock/open and snow-field | Direct runtime observation; lower bound | High | Compared fixed battle frames across the complete user recording | P-VID | Presentation families do not establish authored regions, chapters, mechanical modifiers or the game-wide arena total. |
| C-28 | Ordinary victory signal | Sampled victories follow the opposing living-unit counter reaching zero and then open a result state | Direct runtime observation for sampled battles; not a universal rule proof | High | Followed battle-to-victory-to-report transitions at fixed timestamps | P-VID | Other modes, bosses, timeouts, retreats and defeat terminals were not exhaustively observed. |
| C-29 | Same-battle revival or reserve entry | No qualifying same-battle revival observed; two battle openings expand panels from 2 to 7 and 3 to 7 | Direct observation of staged entry plus difficulty-specific between-battle recovery text | High | Sequentially decoded separate battle openings and reconciled them with visible difficulty rules | P-VID, P-PH3 | Special skills or items may prevent death, but no retained sequence shows a dead member respawning in the same active battle. |
| C-30 | Persistent awards | Distinct observed awards after a 100-victory milestone and after separate 30-, 59- and 126-victory terminals | Direct runtime observation; exact only for each displayed state | High | Read each report and the following persistent-progression handoff without combining values | P-VID, P-PH3 | Currency formulas, version/mode multipliers, upgrade prices and complete reward sources remain unknown. |
| C-31 | Runtime audio mapping in the user recording | Unknown | Unknown because narration and the mixed recording prevent reliable event isolation | High confidence in the unknown status | Sampled battle, shop, victory, reward and final transitions without exporting clips or assigning speculative cues | P-VID | File inventory proves audio assets exist but not their runtime event, mixing, layer or fatigue behavior. |
| C-32 | Observed run-failure terminals | Separate 8-, 9- and 10-member sides reach zero while opponents remain, followed by game over | Direct runtime observation; exact for the three retained terminal sequences | High | Sequentially decoded terminal battles and their game-over/report transitions | P-VID, P-PH3 | Retreat, timeout and special-mode terminals remain unresolved. |
| C-33 | Difficulty-specific death rules | 3 visibly described modes: wipe-and-recover with victory loss; post-victory casualty recovery with party-wipe terminal; permanent-death mode | Direct in-game text observation; exact for the displayed descriptions | High | Retained and sequentially decoded the difficulty screen and battle tutorial | P-PH3 | Settlement details, version drift, retreat outcomes and every special recovery interaction remain unknown. |
| C-34 | Battle execution boundary | Ordinary combat and ultimate activation are automatic; camera, speed, pause and retreat controls are exposed | Direct in-game tutorial observation corroborated by continuous combat | High | Followed tutorial text into active battles without a visible target or ability command | P-PH3 | Exact input grammar and exceptional manual mechanics are not exhaustively tested. |
| C-35 | Formation-card loadout | Catalog surface plus 4 simultaneously populated active slots | Direct runtime observation; exact for one guide state | High | Sequentially decoded hall-to-formation UI and inspected conditional-effect panels | P-PH3 | Total entries, unlocks, mutual exclusions, persistence and effect execution remain unknown. |
| C-36 | Persistent configuration surface | Connected multi-branch Rune Circle plus named save/export/import/load UI | Direct runtime observation for the visible surfaces; persistence is inference only | High for UI, medium for reusable-build inference | Sequentially decoded graph, node details and loadout overlay | P-PH3 | Purchase, prerequisite, refund, import/export and save-survival behavior were not executed. |
| C-37 | Late-run preparation systems | 1 recipe browser layout, separate storage, profession-labeled equipment and 1 completed consumptive passive/veteran operation | Direct runtime observation; lower-bound system coverage | High | Followed source-bound transitions and confirmation/result states | P-PH3 | Complete recipes, pools, probabilities, eligibility and persistence remain unknown. |
| C-38 | Observed final-boss victories | 2 independent ten-member fights reach enemy count 0; one continues to a final-boss decision modal | Direct continuous runtime observation; exact for the two retained fights | High | Sequentially decoded the retained boss, reward and independent final-Raid clips | P-PH3, P-PH4 | Trigger, summon rules, full reward variation and post-retirement state remain unknown. |
| C-39 | Rune spending transaction | Rune-point balance decreases while connected paths visibly illuminate | Direct continuous runtime observation; exact for one retained transaction chain | High | Compared the source-bound before, intermediate and after states | P-PH4 | Refund, reset, save survival and import/export execution remain unknown. |
| C-40 | Advanced-profession gate | One visible node requires 2 adjacent completed paths and a separate currency purchase | Direct UI disclosure; exact for the inspected node | High | Inspected the node before and after adjacent path completion | P-PH4 | The profession purchase itself and every profession's gate were not executed. |
| C-41 | Mobile interaction grammar | Landscape UI with tutorial instructions for Tap reroll, member pickup/swap, item placement and Tap-to-sell | Direct in-game tutorial observation; lower bound for supported operations | High | Reviewed the complete retained mobile walkthrough and state transitions | P-PH4 | Physical coordinates, gesture thresholds, multi-touch, haptics and device ergonomics remain unverified. |
| C-42 | Boss escalation | A seventh boss is visibly bound to round 300, followed by 3 elemental dragon presentations and a separate final Raid | Direct runtime observation; lower bound for the staged boss ladder | High | Followed the retained continuous late-boss sequence and final-Raid entry | P-PH4 | Earlier complete boss list, unlock formulas and cross-version ordering are not exhaustively proven. |
| C-43 | Final-boss continuation choice | Final-boss victory offers continued clan play or voluntary retirement | Direct in-game modal following continuous victory | High | Followed enemy count 0 -> victory -> report -> final-boss modal | P-PH4 | The retirement branch itself and its persistence result were not selected. |
| C-44 | Retirement resource surface | Tavern UI states that battles and retiring members feed a distributable experience resource | Direct UI disclosure only | High for the displayed rule, unknown for execution | Inspected the retained Tavern explanation within a continuous closing segment | P-PH4 | Confirmation, roster removal, equipment settlement, exact conversion and save behavior remain unknown. |
| C-45 | Directly visible version labels | Separate title screens show v1.24.0 and v2.0.0 |
Direct runtime observation; exact for those two sources | High | Read retained title frames without merging their mechanics or values | P-PH3, P-PH4 | Other guide sources have unknown exact builds, and platform/version parity is not assumed. |
Package And Asset Census
These measurements describe the inspected WebGL build. They are included to prevent category errors, not to estimate hidden content.
| ID | Package observation | Count or result | Evidence class and exactness | Confidence | Observation method | Provenance | Entity-count limitation |
|---|---|---|---|---|---|---|---|
| A-01 | Unity objects inspected | 10,879 objects | Direct package observation; exact for the inspected build inventory | High | Parsed the delivered Unity data container with the recorded extraction toolchain | P-EXT, P-MANIFEST | Objects include engine, UI, transforms, assets and components; they are not gameplay entities. |
| A-02 | Readable exported images | 1,638 images: 1,222 Sprite exports plus 416 readable Texture2D exports | Direct package observation; exact for successful exports in the inspected build | High | Counted successful export records by Unity object type | P-EXT, P-IMG, P-MANIFEST | Sprite and Texture2D entries can duplicate the same visual identity or represent layers and frames. A further 25 Texture2D objects had special or empty stream paths and were unreadable. |
| A-03 | Image review categories | 504 animation/VFX, 276 character, 112 enemy, 131 environment, 423 equipment/item and 192 UI/icon images | Direct package observation; exact catalog assignment for A-02 | High for assignment totals; medium for semantic category judgment | Classified every readable image by filename and visual review, then checked that category totals reconcile to A-02 | P-EXT, P-IMG | Categories describe images. Actor layers, faces, equipment overlays, effects, frames and UI are mixed within them. |
| A-04 | Audio clips | 43 exported AudioClip objects | Direct package observation; exact for the inspected build export | High | Counted successful AudioClip exports | P-EXT, P-MANIFEST | A clip may serve multiple events, and an event may layer several clips; this is not an event or content count. |
| A-05 | Animation metadata objects | 208 AnimationClip objects | Direct package observation; exact for the inspected build inventory | High | Counted AnimationClip metadata objects | P-EXT, P-MANIFEST | Clips can be shared, layered, transitional or unused; they do not equal attacks, skills or entities. |
| A-06 | Script type identities | 2,195 MonoScript records, including 196 project/game assembly types | Direct package observation; exact for the inspected build inventory | High | Enumerated type identities from Unity metadata and assembly membership | P-CODE, P-EXT | Type identity proves an architectural surface, not record totals, formulas, method bodies or player-visible features. Recovered original C# source count is exactly 0 for this inspection. |
| A-07 | Character face families | 26 asset-name families | Inference from direct image catalog inspection; exact for the catalog grouping, not a character count | Medium | Grouped recurring face asset names and checked visual layers | P-AUDIT, P-IMG | Faces can be variants, reusable layers or procedural options; they cannot be promoted to named recruits. |
| A-08 | Gameplay database serialization | Inspected components are 32-byte shells with no exposed custom arrays | Direct package observation; exact for the inspected component instances | High | Read the available MonoBehaviour type trees for recruit, enemy, item, artifact and progression database surfaces | P-AUDIT, P-CODE | Runtime initialization occurs inside compiled IL2CPP WebAssembly, so shell size and type existence reveal no entity totals. |
Reconciliation Rules
The following equivalences are explicitly invalid:
| Observed evidence | It must not be reported as | Safe use |
|---|---|---|
| A-02 and A-03 image counts | Total characters, enemies, items, skills, biomes or UI screens | Asset-production scale and category coverage only |
| A-03 enemy-category images | Enemy identities | Evidence that enemy presentation is layered and has production breadth |
| A-03 equipment/item images | Item records | Evidence that items require icons, overlays, rarity treatment and UI states |
| A-07 face families | Named recruits | Evidence for modular portrait or actor customization scope only |
| C-06 visible slot/icon identities | Active skill definitions | Evidence for visible progression density only |
| A-05 animation objects | Abilities or combat actions | Animation-production scope only |
| A-04 audio objects | SFX events, music states or gameplay entities | Audio-production scope only |
| C-16 environment asset families | Chapters, biomes traversed or encounter waves | Environment-production lower bound only |
| A-06 type identities | Implemented features or content rows | Architecture vocabulary and research routing only |
Why Exact Totals Cannot Be Recovered
The delivered program uses compiled IL2CPP WebAssembly plus runtime metadata. Type identities and component shells are visible, but original method bodies, runtime-initialized arrays and a complete player-facing collection are not. The database components inspected for recruits, enemies, items, artifacts and progression contain no readable custom arrays in the available type tree.
Consequently, exact totals for recruits, enemy identities, items, active skills, artifacts, authored formations, chapters, waves, achievements, collection entries and modes remain unknown. Inferring those totals from filenames, frames, body parts, rarity layers, component instances or script types would be false precision.
Defensible Census Summary
The archive directly proves a run-based preparation and auto-battle loop with multiple recruit labels, base and advanced profession identities, a branched persistent progression surface, several ordinary enemy roles, equipment, milestone artifacts and editable party order. It also proves that a completed captured adventure reaches a finite result and feeds persistent progression.
The strongest defensible numerical statements are the lower bounds and scope-limited exact counts in C-01 through C-32. No row should be read as a complete current-version encyclopedia unless its exactness cell explicitly states that scope.
Remaining Authorized Census Work
An exact gameplay census would require at least one of the following authorized surfaces:
- complete in-game indexes for roster, professions, enemies, items, artifacts, skills, achievements and collections;
- a complete campaign/chapter selector with wave or encounter accounting;
- an owner-authorized runtime data export that exposes record arrays without bypassing access controls.
Until such evidence exists, the unknown rows above are the correct terminal result, not incomplete arithmetic.