ark-stronghold-protocol/docs/DATA.md

DATA.md — generated game data (data/*.json)

All files in data/ except data/assets.json are produced by node tools/build-data.mjs (task F1) from the official zh_CN client data (Kengxxiao/ArknightsGameData) joined with docs/research/*.json. Do not edit them by hand — change the build script and rebuild.

node tools/build-data.mjs              # build (downloads missing official files into .cache/gamedata/)
node tools/build-data.mjs --offline    # never download; fail if a cached file is missing
node tools/build-data.mjs --refresh    # re-download every official file
node tools/build-data.mjs --no-research --out /tmp/x   # research-free build (fallback defaults), other dir
node tools/build-data.mjs --cache <dir> --report <file> # other cache dir / report file (tests, CI)
node tools/build-data.mjs --force      # write data/ even when integrity checks fail (debugging only)
node --test test/data.test.js          # integrity tests (+ raw-data cross-check and rebuild check when .cache/gamedata exists)

Unknown options or a missing option value are errors (exit code 2); --refresh and --offline are exclusive.

• Cache. Official files live under .cache/gamedata/<repo path> (excel/activity_table.json, levels/enemydata/enemy_database.json, levels/activities/act1autochess/level_*.json, …) and are fetched from https://raw.githubusercontent.com/Kengxxiao/ArknightsGameData/master/zh_CN/gamedata/<path> when missing. Downloads are validated (must parse as JSON) and written atomically.
• Report. Every run writes .cache/build-data-report.json (counts, byte sizes, warnings (each once), errors, written). Integrity errors (see §17) make the exit code 1 and leave the previous output untouched (unless --force); warnings never do. Each output file is written atomically (temp file + rename).
• Determinism. Same inputs ⇒ byte-identical outputs (stable key order, no timestamps, no randomness).
• Size. ≈3.5 MB total (limit 6 MB; chess.json ≈1.65 MB with the loadout choices), compact JSON (no indentation).
• Derived paths. stages.json groundPaths* come from the sim's own server/sim/grid.js pathing: a change there needs a rebuild (the offline-rebuild test catches a stale data/).

0. Conventions used by every file

Convention	Meaning
ids	Always the official ids: chess_char_1_01_a (normal) / _b (精锐 golden), chess_item_1_01_e_a, garrison_16_a, bondeffect_yan, enemy_1422_lrsldr, band_bldsk, boss_1, act1autochess_03 (wave template = level file name, lower-case), act2autochess_m01 (stage).
desc / descRaw	descRaw keeps the official rich-text markup (<@ba.vup>+15%</>, <$ba.sluggish>停顿</>), desc is plain text. Literal trigger labels such as <获得时>, <在场6名不同【炎】干员> are part of the text and kept in both. \n escapes are real newlines. Placeholders ({atk:0%}, {-x}, {x:0.0}) are already resolved at the record's level, except bond effectDesc (see bonds).
bb / bbStr	Flattened official blackboard. bb[key] = number; bbStr[key] = the official valueStr when present. A string-valued entry (value 0 + valueStr) appears only in bbStr. Duplicate keys get _1, _2 suffixes. Numbers are float-noise-cleaned (6 decimals).
params	Effects/items/bands: {…bb, …bbStr} of all buffs merged, first occurrence wins (convenience; use buffs for exact semantics).
positions	[row, col], row 0 = bottom, col 0 = left, 19 × 21 grid (DESIGN §3).
rangeGrid	[[dRow, dCol], …] relative to the unit facing right ([0,0] = own tile). Rotated for the unit's direction (DESIGN §3, server/sim/dir.js rotateOffset): RIGHT (dr,dc), UP (dc,−dr), LEFT (−dr,−dc), DOWN (−dc,dr); rangeExtend grows +dCol before rotating. (Every grid in the data is symmetric top-to-bottom, so LEFT equals mirroring dCol.)
round-keyed maps	JSON object keys are strings: rounds["14"].
[ASSUMED] / assumed: true	Value is a documented reconstruction (research), not client data. Keep it tunable.

---

1. config.json — modes, economy and all global tunables

Top level: { season, seasonName, modes, economy, lpCapPerRound, bossOvertimeAfter, bossOvertimeDrainPerSec, bossHpScale, hiddenCore, dp, unite, finalAssault, timers, bans, bandDraft, titles, titleRule, tips, broadcasts, trophies, roundScores, rewards, constants }.

1.1 modes[modeId] (9 modes; modeId = mode_${single|multi}_${difficulty} + mode_training_1)

Field	Example	Meaning
modeId, name, code, sortId	"mode_multi_abyss", "终极模拟", "AC-4", 5	official
type	"MULTI"	SINGLE 独立 / MULTI 同盟 / LOCAL training
difficulty	"ABYSS"	FUNNY NORMAL HARD ABYSS (TRAINING)
inScope	true	false for training (out of scope v1)
color, iconId, backgroundId	"#ff0024", "mode_abyss_icon"	UI
desc, effectDescList, unlockText		official texts
specialPhaseTime	16	official 机变 phase time field (multi 16 s per picker, solo 150)
activeBondIds / inactiveBondIds	FUNNY: 13 / 10	static bond availability
inactiveEnemyKeys	18 keys (FUNNY/NORMAL)	enemies never used in this mode
lastRound	14 (solo FUNNY 9)	last regular round (boss round)
bossRound	14	round of the Final Assault
hiddenRound	15 or null	隐秘核心 round (NORMAL+)
rounds[r]	see below	per-round schedule
spRounds	[3,9,11]	rounds whose prep opens with a 机变 draft
combatTimeLimit[r]	{"1":45,…,"14":null}	= rounds[r].combatTimeLimit: real seconds of the forced-2× battle (DESIGN §4)
enemyScale[r]	{"atk":1.1,"hp":1.2,"speed":1,"kAtk":1,"kHp":1}	non-boss enemy multipliers (research 01 A3): atk = atkBase·1.1^kAtk, hp = hpBase·1.2^kHp·extra, speed 1.15 on ABYSS from R3. Apply to level-0 stats after special-enemy replacement, also to bounty/special enemies. Leader HP pools excluded.
bossHpScale	multi {"bloodPointKey":"bloodPointAbyss","coop":1,"aliveScaling":false,…} · solo {"bloodPointKey":"bloodPoint","solo":0.25,"soloAssumed":true,…}	co-op pool = bosses[id].bloodPoint[difficulty] unscaled (split leaders share one pool, "敌方领袖的总生命值不变", no alive-player factor — research 08 §6 #7); solo = × solo [ASSUMED, flagged]. GameData.bossPoolHp(bossId) implements it
upgradePrices	[5,8,11,12,13]	base price L1→2 … L5→6 (−1 per round start, floor 0, reset after upgrade)
maxShopLevel	6
shopSlots[level]	{"1":{"chess":3,"item":1},…}	operator + item slots per shop level
levelTagColors[level]	"#434343"	UI
stages	["act1autochess_m02",…]	active stages allowed (pick by stages[id].weight, 50 each): 标准 FUNNY (and training) = act1autochess_m01 战场#01 only; 险境 NORMAL = the 8 active stages; 绝境 HARD / 终极 ABYSS = 7 (no 战场#01) — official mode data
bossWeights / hiddenBossWeights	{"boss_1":6,…} / {"boss_8":50,…}	weighted boss picks

rounds[r]:

Field	Example	Meaning
template	"act1autochess_03"	wave template (waves.json) for non-boss rounds; null for boss rounds
bossTemplates	{"boss_1":"act1autochess_h07_01",…}	boss rounds: template per bossId (solo uses _s)
combatTimeLimit	55	real seconds (= template maxPlayTime) of the forced-2× battle: the Battle / 联防 limit is 2 × this in game seconds (gamedata.js combatTimeLimit, config.combatTimeScale default 2; docs/BALANCE.md §2.1); null for boss rounds (no hard stop)
levelMaxPlayTime	120	raw maxPlayTime of the (first) template in real seconds; boss rounds: the Final Assault / Hidden Core countdown (m.public.deadline, gamedata.js bossLevelTime) — not a hard stop, the battle goes on
prepTime	65	multi/training prep timer (real seconds); null for solo (untimed in 下半)
prepTimeData	300	official normalPhaseTime (optional "timed solo")
isSpPrepare, isBoss, isHidden		机变 before prep / leader round / 隐秘核心
bossOvertimeAfter	150	boss rounds: team LP −1 per real second after this many real seconds (= 300 game s on the 2× field clock; m.public.overtimeAt)

1.2 economy

Field	Example	Meaning
income	[0,4,5,6,…,12]	index = round (index 0 unused): min(3+r, 12); incomeCap 12, incomeCapAlternative 10, incomeAssumedAfterRound 3
leftoverFundsLost, leftoverFundsKeptByBands	true, ["band_cannot"]
chessPrice[tier] / chessSell[tier]	{"normal":2,"golden":2} / {"normal":1,"golden":1}
chessStatus[tier]	{"normal":{"phase":1,"level":55,"skillLevel":4,"equipLevel":0,…},"golden":{…}}	training status per tier
refreshPrice	1
freeze	{"scope":"allUnsoldSlots","price":0,"consumedAtRoundStart":true,"refreshWhileFrozenRerollsAll":true,…}	one toggle freezes all unsold slots
shopClearedAtCombatStart	"unfrozenSlots"
itemSellable, itemDestroyRefund	false, 0	items can only be destroyed
benchSize, tempSize, deployCap, storeCntMax	10, 5, 8, 6
equipPerChess, maxArtsPerRound	2, 2
poolCopies[tier] + poolCopiesOverrides	{"1":12,"2":14,"3":18,"4":16,"5":8,"6":5}, {"chess_char_6_11_a":4} (缪尔赛思)	shared pool copies per base chess
goldenCopies, mergeCount, mergeCountOverrides, itemMergeCount	3, 3, {"chess_char_2_11_a":2} (风丸), 2
rewardOffer	{"count":3,"tierOffset":1,"maxTier":6,"price":0,…}	merge reward: 3 free chess of tier min(shopLevel+1,6)
handFillOrder	"rightToLeft"
shopOdds	{"model":"copyWeighted",…}	[ASSUMED] roll model
borrowCount, fallbackBondId, defaultBandId, defaultStartLp	20, "emptyShip", "band_bldsk", 28

1.3 Other global keys

Key	Example	Meaning
lpCapPerRound	10	max LP lost per player per normal round
bossOvertimeAfter, bossOvertimeDrainPerSec	150, 1	official bossTurnHpReduceTime: from 150 real s of a boss round the merged team LP loses 1 per whole real second (gamedata.js bossOvertimeDue; first point at 151 s)
bossHpScale	{"formula":"co-op: bloodPoint[difficulty] …","coop":1,"solo":0.25,"soloAssumed":true,"aliveScaling":false,…}	global form of the mode rule
hiddenCore	{"single":350,"multi":1200,"minTeamLpExclusive":1,"difficulties":["NORMAL","HARD","ABYSS"],"checkedAfterRound":14}	Σ activated layers must be > threshold and LP > 1
dp	{"init":10,"perSec":1,"max":99}
unite	{"maxHelpers":2,"helperOrder":"unitsOnField>activeBond>undownedUnits; pair: unitsOnField>activeBond>activeLayers>undownedUnits, first = right field (PRTS 帮助)","layerGainsEnabled":false,"templates":{"1":"act1autochess_escaped_single","2":"act1autochess_escaped_multi"},…}	联防; helperOrder documents the rule server/match/unite.js helperOrder implements (research 08 §5; ties → seat)
finalAssault	{"pairing":"seatOrderPairs","oddPlayerAlone":true,"movableBossPerAlivePlayerSide":true,"layerGainsEnabled":false}
timers	{"infoCheck":25,"infoCheckHint":5,"bandDraft":50,"bandDraftHint":15,"bandTurn":12,"battleCheck":3,"spFirst":30,"spTurn":16,"soloPrepTimeData":300,"soloSpTimeData":150,"chatCd":1,"chatBubble":3,"broadcastDelay":1,"enterSteps":[…]}	seconds
bans[difficulty]	{"core":3,"addon":4} (FUNNY {0,1}) + rule	per-match disabled-bond draw (research 01 A2)
bandDraft	{"skipsPerPlayer":1,"order":"random","duplicatesAllowed":true,"timeoutBandId":"band_bldsk"}
titles[]	{"id":"comment_1","picId":"comment_icon_1","name":"卫戍之星","stat":"bossDamage","rule":"max","onlyOnWin":true,"text":"…"}	评语 + [ASSUMED] criteria; titleRule explains assignment
tips[]	{"tip":"只有在联合模拟中才能获得奖杯","weight":50}	gameTipsList
broadcasts[]	{"id":"comment_boss_hit_1","type":"BOSS_HIT","priority":30,"text":"{0}博士对敌方领袖造成的伤害超过20%!","textRaw":"<@ba.vup>{0}博士</>…","params":["0.2"]}	{0} player, {1} operator/level, {2} amount
trophies	{"byRoundsPassed":[{"maxRound":4,"FUNNY":0,…},…],"hiddenCore":{…},"multiOnly":true,"medals":[{"count":0,"iconId":"trophy_level1_icon"},…]}
roundScores[]	{"round":4,"score":50}
rewards	{"itemId":"act2autochess_token_chess","baseByRoundsPassed":[{"round":1,"count":10,"dailyPoint":10},…],"formula":"…","difficultyFactor":{…},"modeFactor":{…}}	卫戍认证
constants	{"maxLevelCnt":15,"bossTrailerStartRound":3,"singleReconnectTime":86400,…,"pingConds":[…]}	misc official constData; singleReconnectTime (s) = how long a dropped solo run stays resumable (server/lobby.js; co-op keeps the 10-minute window)

---

2. chess.json — { [chessId]: Chess } (266 = 133 normal + 133 golden)

112 are visible (non-hidden, non-DIY) normal chess: per tier 16/17/19/22/19/19. 17 are isHidden (retired 上半 entries or effect-only such as chess_char_1_15_a 盟约·辅助干员 from band Pith); 4 are DIY (甄选) slots.

Field	Example	Meaning
chessId, baseId, goldenId, isGolden	"chess_char_1_01_b", "chess_char_1_01_a", "chess_char_1_01_b", true	normal/golden pair
tier	1	阶 I–VI (shop tier)
identifier, shopSortId	133, 1	official ordering
isHidden, isDiy, visible		visible = !isHidden && !isDiy (only visible chess enter the shop pool)
chessType	"PRESET"	PRESET / NORMAL / DIY
charId	"char_498_inside"	operator (null for DIY)
name, appellation	"隐现", "Insider"
rarity	5	stars 1–6
profession, subProfessionId, subProfessionName, position	"SNIPER", "fastshot", "速射手", "RANGED"
nationId	"laterano"
bonds[]	["lateranoShip","swiftShip"]	bondIds (→ bonds.json)
garrisonIds[]	["garrison_16_b"]	特质 (→ garrisons.json); first = displayed trait
price, sellPrice	2, 1
upgradeNum, upgradeChessId	3, "chess_char_1_01_b"	copies needed to merge (风丸 2; golden 0)
status	{"phase":2,"level":50,"skillLevel":7,"equipLevel":1}	training status used for all numbers
stats	{"maxHp":1421,"atk":508,"def":167,"res":0,"cost":15,"blockCnt":1,"bat":1,"aspd":100,"respawnTime":80,"spRecovery":1,"hpRecoveryPerSec":0,"moveSpeed":1,"tauntLevel":0,"massLevel":0,"deployLimit":1,"deckStack":0}	keyframes linearly interpolated at status, hp/atk/def/cost/block/respawn rounded; golden includes the module attribute bonus. aspd 100 = base (module attack_speed added). Talent stat boosts are NOT folded in.
immunities	{"stun":false,"silence":false,"sleep":false,"frozen":false,"levitate":false}
rangeId, rangeGrid	"3-3", [[1,0],[1,1],…]	attack range at status
dmgType	"phys"	phys / arts / heal (derived, see §2.1)
attackKind	"ranged"	melee / ranged / heal / none
projectile	"arrow"	arrow (phys ranged) / bolt (arts ranged) / orb (heal) / none
canHitFly	true
targetPriority	"fly"	from trait text: fly (优先攻击空中单位), lowestDef (防御力最低), else null
trait	{"desc":"优先攻击空中单位","descRaw":"…","bb":{"atk_scale":1.1},"bbStr":{},"rangeGrid":null,"moduleDesc":"攻击空中单位时攻击力提升至110%","moduleDescRaw":"…"}	profession trait (+ golden module trait upgrade merged into bb; rangeGrid = trait-effect area, e.g. 散射手 front row — not the attack range)
skill	see below	default skill at status.skillLevel (normal 4, golden 7)
skills[]	[{…skill record…, "index":0, "isDefault":false}, {…, "index":1, "isDefault":true}]	loadout choices (DESIGN §16): every skill unlocked at status (E1 ⇒ S1–S2, E2 ⇒ S1–S3; the default is always listed), same shape as skill + isDefault, at the chess skill level, trigger resolved for that skill index (§2.2). The isDefault entry equals skill
modules[], statsBase, traitBase, talentsBase	see §2.2	golden chess with equipLevel > 0 only: selectable modules + the no-module base they apply to
talents[]	{"index":0,"name":"火力支援","desc":"…","descRaw":"…","bb":{"self_ammo":3,"duration":20,"ally_ammo":1},"bbStr":{},"rangeGrid":null,"tokenKey":null,"hidden":false,"fromModule":false}	best unlocked candidate at status (potential 0); golden: module talent upgrades applied (fromModule); module data-only talents have name:null, hidden:true. A module upgrade of an existing talent merges blackboards (module keys win, base keys it does not restate are kept — e.g. 宴 keeps min_attack_speed; 仇白's upgrade adds atk_scale_t next to the old atk_scale: prefer the key the text uses). Module parts flagged isToken are not applied to the operator; they upgrade its summons (tokens.json variants). containerTokenKey: the official talent token id when it is a container missing from character_table (凛御银灰), tokenKey then holds the default skill's token
tokens[]	["token_10028_vigil_wolf"]	summons (→ tokens.json): displayTokenDict + default-skill overrideTokenKey + talent tokenKey; tokens.json → variants[chessId].sources tells which (a display-only token is not produced by this chess's default skill or talents)
module	{"id":"uniequip_002_inside","name":"“最初的惊喜”","type":"MAR-X","level":1,"active":true}	active only on golden chess
assets	{"avatar":"char_498_inside_2","portrait":"char_498_inside_2","spine":"char_498_inside","skillIcon":"skchr_inside_2","subProfIcon":"sub_fastshot_icon"}	asset ids (URLs in data/assets.json); golden uses the E2 art when it exists
diyRequirement	"TIER_6"	DIY only

skill:

Field	Example	Meaning
skillId, iconId, name, level, index	"skchr_inside_2", "skchr_inside_2", "“解决麻烦”", 7, 1	index = character skill slot
desc, descRaw	"攻击力+100%，…攻击装有14发弹药…"	placeholders resolved
skillType	"MANUAL"	MANUAL / AUTO / PASSIVE
durationType	"AMMO"	NONE / AMMO
duration	0	seconds (0 = instant/ammo)
spType	"INCREASE_WITH_TIME"	also INCREASE_WHEN_ATTACK, INCREASE_WHEN_TAKEN_DAMAGE, ON_DEPLOY (passive, no SP)
spCost, initSp, maxChargeTime, increment	20, 10, 1, 1	maxChargeTime > 1 ⇒ charges
bb, bbStr	{"atk":1,"base_attack_time":-0.3,"attack@trigger_time":14}	kit input (DESIGN §5.6)
rangeId, rangeGrid	null	skill range override
prefabId, overrideTokenKey
trigger	{"rule":"DEFAULT","rawRule":"DEFAULT","customRangeGrid":null}	auto-cast rule. rule ∈ DEFAULT, TAKE_DAMAGE (古米/灰毫 S1 only), SP_FULL (official ALWAYS), CUSTOM_RANGE (official CUSTOM_RANGE_SEARCH_ENEMY, uses customRangeGrid), SEARCH (安洁莉娜), MLYSS_WTRMAN (缪尔赛思), GDGLOW_SKILL_2 (荒芜拉普兰德, 纯烬艾雅法拉). Resolution per research 03 Addendum C1.

2.1 Combat classification heuristic (dmgType / attackKind / projectile / canHitFly)

• dmgType: MEDIC (except incantationmedic) and bard → heal; trait text containing 法术伤害 or profession CASTER → arts; else phys.
• attackKind: bard, phalanx, librator → none (no normal attack); heal → heal; position RANGED or melee sub-professions with a ranged normal attack (lord, fortress, shotprotector, agent, hookmaster) → ranged; else melee.
• canHitFly: ranged attackers unless the trait says 地面敌人 (投掷手); skywalker (蒂比) true.
• These are defaults for the generic engine; kits may override.

2.2 Loadout choices: skills & modules (DESIGN §16)

Before a match the player may pick, per base chess, the equipped skill (any index legal for both the normal and the elite record — shared/protocol.js loadoutOptions) and, for the elite, the module (or none). The data carries everything needed to resolve a unit for (chessId, skillIndex, moduleId) (simdata resolveLoadout / loadoutRecord, DataSource.getChess(id, loadout)):

• skills[] — see the table above. Triggers per skill: charId rows (exact skillIndex, or −1 = every skill) → subProfession rows → profession rows with that exact index (official skillIndex 0 = skill 1 only: every defender's S1 is TAKE_DAMAGE, their S2/S3 DEFAULT); customRangeGrid from skillRangeDict[skillId].
• modules[] (golden, equipLevel > 0): every ADVANCED uniequip of the character (the INITIAL uniequip_001_* is "no module" = choice 'none'), in official order, at the chess equipLevel (1, T6: 3):

Field	Example (chess_char_6_11_b)	Meaning
uniEquipId, name, typeName, typeIcon, icon	"uniequip_003_mlyss", "落叶四季", "TAC-Y", "tac-y", "uniequip_003_mlyss"
isDefault, level	false, 3	default = defaultUniEquipId (= module.id)
attr	{"maxHp":170,"atk":28,"def":28}	flat stat additions (stat field names): stats = statsBase[f] + attr[f] (float-noise cleaned)
traitOverride	{"desc":"…提升至165%","descRaw":…,"bb":{"atk_scale":1.65},"bbStr":{},"rangeGrid":null,"moduleDesc"?:…} | null	the full trait with this module; null ⇒ traitBase
talentChanges[]	{"talentIndex":1,"name":"开源节流","desc":…,"descRaw":…,"bb":{"cost":-2,"runtime_cost":-1},"bbStr":{},"rangeGrid":null,"tokenKey":null,"hidden":false}	talent additions/overrides (talentIndex −1 = new hidden data-only talent); applied to talentsBase with the build's merge rule (simdata composeTalents: override of an existing index merges blackboards, module keys win)

• statsBase / traitBase / talentsBase — the golden record without any module (stats / trait / talents keep the default module, unchanged). test/data.test.js proves that composing the default module onto the base reproduces stats, trait and talents exactly for every golden chess.
• Module-less goldens (蒂比, 凛御银灰): modules: [], base = own values. Normal chess have none of these fields.
• Module parts flagged isToken and the per-token attribute blackboards go to the summons: tokens.json → variants[owner].byModule (§14). Module choices never change the combat classification (dmgType…; the build warns if one ever would).

---

3. bonds.json — { [bondId]: Bond } (23 bonds, 8 core)

Field	Example (deputShip)	Meaning
bondId, name, identifier	"deputShip", "助力", 14
isCore, bondType, bondOrder, powerIdList	false, "REGULAR", 2, []	core = isPower (SEASON); powerIdList = nation/group ids for DIY bond derivation
iconId	"icon_deputShip"
activeCount	2	official activation count
thresholds	[2,3]	ascending member counts that raise the tier (tier = number of thresholds reached). yan [3,6,9], egir [3,5], sunt [2,5], solo [1]
maxCount	null (solo 1)	count_threshold_downward: active only while count ≤ maxCount
thresholdTemplate	"count_threshold_upward"	also _downward (独行), _upward_golden (绝技)
countMode	"BOARD"	BOARD (distinct base chess on board), BOARD_AND_DECK (+hand: 远见/奇迹/投资人), BOARD_ALL_CHESS (绝技: every golden chess on board, duplicates count)
countsHand, countsGoldenOnly	false, false	convenience flags
activeType, isActiveInDeck, noStack, weight, maxInactiveBondCount	"BATTLE", false, false, 10, -1	weight 0 ⇒ never drawn for per-match bans
layerMilestones[]	[{"layer":25,"mode":"every","effect":"bond_layer_added_reward_equip"}]	layer-based powers: reach (while L ≥ layer), every (each multiple), first (latched once)
desc, descRaw		bond panel text
effectId, effectName, effectDesc, effectDescRaw	"bondeffect_deput", …, "所有干员防御力+{0:0%}（受层数影响）…"	in-battle text with positional placeholders
effectDescParams[]	[{"index":0,"format":"0%","base":"base_def","perStack":"def_per_stack"}]	{i:fmt} = bb[base] + bb[perStack] × layers
bb, bbStr	{"base_def":0.15,"def_per_stack":0.012,"respawn_time":-0.3,"layer":2,"count":3,"more_layer":4}	all buffs merged (first wins)
buffs[]	[{"key":"env_gbuff_new","bb":{…},"bbStr":{"key":"act1autochess_bond_eff_deput"}}, …]	exact official buffs (e.g. 萨尔贡 has a second buff valid only with band_narant)
baseParams, perStackParams	["base_def"], ["def_per_stack"]	official descParamBaseList / descParamPerStackList
members[], visibleMembers[]	["chess_char_1_06_a",…]	base (normal) chessIds incl. hidden / only visible
spec	{"tiers":[{"need":"…","effect":"…"}],"formulas":{…},"layerGain":"…",…}	research 02 implementer spec (null without research)

4. garrisons.json — { [garrisonId]: Garrison } (249; 43 distinct effectKeys)

Field	Example	Meaning
garrisonId	"garrison_30_a"	_a normal, _b golden version
desc, descRaw	"<获得时>获得等于当前调度中心等级的【炎】层数（无需激活盟约）"
eventType, eventTypeDesc, eventTypeIcon	"SERVER_GAIN", "单次叠加", "icon_bond"	IN_BATTLE, SERVER_GAIN 获得时, SERVER_PREP_START 进入休整期时, SERVER_PREP_FIN 休整期结束时, SERVER_CHESS_SOLD, SERVER_PRICE, SERVER_REFRESH_SHOP
effectType	"SERVER_ADD_BOND_METHOD"
effectKey	"SERVER_ADD_BOND_METHOD" / "act1autochess_gar_eff_attrByBond"	handler key: bbStr.key (battle runes) else effectType
battleRuneKey, charLevel	null, 0
bb, bbStr	{"multi":1}, {"add_method":"shoplv","bond":"yanShip"}
owners[]	["chess_char_1_03_a"]	chess carrying it (includes garrisons referenced by other garrisons, which may have no owner)

5. items.json — { [itemChessId]: Item } (115 = 56 EQUIP × 2 + 3 MAGIC)

Field	Example (chess_item_1_01_e_a)	Meaning
id, baseId, goldenId, isGolden	"chess_item_1_01_e_a", …, "chess_item_1_01_e_b", false
trapId, iconId	"trap_1041_acarm041"	icon id = trapId
identifier, shopSortId	267, 1
name	"维式重锤"
itemType	"EQUIP"	EQUIP (equipped, max 2) / MAGIC (Arts, dropped on a tile)
tier, price, hideInShop	1, 1, false	golden price 5 (never sold in shop)
mergeable, upgradeNum, upgradeChessId	true, 2, "chess_item_1_01_e_b"	2 copies → golden; upgradeNum 100 = never merges
duration	-1	trapDuration
giveBondId, givePowerId, canGiveBond	"victoriaShip", null, false	变形同构体 (canGiveBond) grants the giveBondId of the other equipped item
requiresBondId	null / "yanShip"	VI-tier bond signature items (research 04)
effectId, effectName, desc, descRaw	"eff_acarm041", …, "攻击力+15%"
buffs[], params	[{"key":"env_gbuff_new_with_verify","countType":"NONE","bb":{"atk":0.15},"bbStr":{"key":"attr_common_global_buff"}}]
category, kind, family, implFormula	"STAT", "passive", null, "ATK% += atk …"	research 04 (categories: STAT, ON_HIT, SURVIVAL, SP, ECONOMY, RECRUIT, BOND, BOND_GRANT, SET, BOND_SIGNATURE, MAGIC)
rangeGrid	[[0,0]]	Arts area (画卷 [[0,0],[0,1]])
flavor

6. bands.json — { [bandId]: Band } (40)

Field	Example	Meaning
bandId, sortId, name, iconId	"band_bldsk", 1, "华法琳", "icon_bldsk"
modeTypeList	["LOCAL","SINGLE","MULTI"]
totalHp	28	starting LP (20–45)
effectId, effectName, desc, descRaw	"aceffect_band_21", "重点监护", "【重点监护】开始作战时…"
buffs[], params	[{"key":"prep_finish_char_bond_add_layer","bb":{"layer":2},"bbStr":{}}]
victorCount, rewardModulus, unlockDesc	3, 1, null

7. effects.json — { [effectId]: Effect } (361: every effect of the season)

Types: BAND_INITIAL 41, BOND 23, EQUIP 115, ENEMY_GAIN 129 (悬赏), BUFF_GAIN 43 (战术决策), ENEMY 6 (aceffect_enemy_* global modifiers), CHAR_MAP 4.

Field	Example	Meaning
effectId, effectType, name	"aceffect_band_21", "BAND_INITIAL", "重点监护"
desc, descRaw
counterType, continuedRound, decoIconId, enemyPrice	"NONE", -1, null, 0	enemyPrice = bounty coins
buffs[]	{"key":"…","countType":"NONE","bb":{…},"bbStr":{…}}
params	{"layer":2}	merged convenience view

8. choices.json — 机变 draft data

Key	Example	Meaning
events[id]	{"id":"enemy_initial_1","choiceType":"BOUNTY_HUNT","effectType":"ENEMY_GAIN","name":"悬赏决策","desc":"…","color":"#35d8b4","family":"bounty","solo":false,"training":false}	109 official choice events; _s = solo copies; usedBy on events opened by items (hunter_band_1 ← 教鞭 / “神秘顾客”)
families	{"bounty":{"name":"悬赏决策",…},"supply":{…},"shop":{…},"tactic":{…}}	悬赏决策 / 道具补给 / 机密商店 / 战术决策
format	{"multi":{"cards":6,"pickOrder":"random","firstPickSec":30,"otherPickSec":16,"onTimeout":"autoPickRandom","eachPlayerPicks":1},"solo":{"cards":3,"timer":null},"opensAfterIncome":true}
cards.bounty[]	{"effectId":"enemyeffect_10_4","name":"悬赏·飞行I","tier":1,"coin":1,"payout":"kill","rounds":2,"multiRound":false,"enemyKey":"enemy_1005_yokai","count":1,"adds":[…]}	payout kill (killer gets coins, 联防 helpers too) / perfect (chooser's own phase perfect); rounds = battles affected (99 = all remaining); tier from the I/II/III suffix (else coin value)
cards.tactic[]	{"effectId":"map_m01_1","name":"模拟战场演变·模式一","kind":"terrain","stageId":"act1autochess_m01","team":false}	kind ally / enemyDebuff / terrain (only for the match stage); team = also given to teammates
schedule[modeId]	{"spRounds":[3,6,9],"rounds":{"3":{"families":[{"family":"bounty","weight":50},…],"cards":6,"supplyTiers":[1,4],"bountyTiers":[1,2],"events":{"bounty":[…],"supply":[…],"shop":[…],"tactic":[…]},"assumed":true}}}	research 01 A4 defaults: pick a family by weight; supplyTiers = item tier window for 道具补给; 机密商店 draws any tier; events = official event ids usable as the phase header
pools[poolId]	{"kind":"equip","items":[…]} / {"kind":"chess","tier":2,"rule":"shopEligible"}	server-side reward pools referenced by effects (pool_equip_*, pool_chess_shop_N_reward, pool_char_later, …), all [ASSUMED]; rule:"shopEligible" = visible/non-hidden entries, maxTier:"shopLevel" = ≤ current shop level, weighted = [[id, weight]]

9. enemies.json — { [enemyKey]: Enemy } (249 keys that can appear)

Collected from all wave templates (spawns, branches, enemyDbRefs), special-enemy entries and pools, bounty effects, band effects, bosses, placeholders, mode bans, plus summons (extraEnemyKeyList) and enemy keys in enemy blackboards (transitively). Level = randomEnemyAttributeDict[key].level (0 for all); the season override level level_autochess_enemy_data is applied (灼藤/元核孽生者 ATK 400, …).

Field	Example (enemy_1422_lrsldr)	Meaning
key, name, level, rank, handbookIndex	…, "萨卡兹枯朽前锋", 0, "NORMAL", "NAC18"	rank NORMAL / ELITE / BOSS
desc, descRaw, applyWay	…, "MELEE"	applyWay MELEE / RANGED / ALL / NONE
stats	{"maxHp":4700,"atk":300,"def":100,"res":40,"moveSpeed":0.8,"bat":3,"aspd":100,"rangeRadius":0,"rawRangeRadius":-1,"blockCnt":1,"massLevel":1,"lpr":1,"hpRecoveryPerSec":0,"elementRes":0,"elementDmgRes":0,"hitRatePhys":0,"hitRateArts":0,"dmgType":"phys","dmgTypes":["phys","arts"],"motion":"WALK","immunities":{…},"otherImmunities":[],"tauntLevel":0}	level-0 (unscaled). rangeRadius ≥ 0 is the normal-attack radius under the DESIGN §5.5 contract: > 0 ⇔ the enemy attacks units it is not blocked by; always 0 for applyWay MELEE (melee enemies only hit their blocker; 粉碎攻坚手/宿主士兵/深池方阵步兵 carry an official 1.2–2.5 used by abilities) and negative sentinels become 0. rawRangeRadius = the official value. blockCnt = block slots it consumes; lpr = lifePointReduce; dmgType = first handbook damage type (phys/arts/heal/none/…), full list in dmgTypes. Undefined database fields (m_defined:false) holding the zero value fall back to the defaults lpr 1, bat 1, aspd 100, moveSpeed 1 (see §15)
abilities[]	{"text":"位于源石污染区内时，攻击造成法术伤害","textRaw":"…","format":"NORMAL"}	handbook ability list
talents	{"bb":{"1.hp_ratio":0.2,"1.damage_scale":0.5},"bbStr":{}}	talentBlackboard
skills[]	{"prefabKey":"2","priority":0,"cooldown":50,"initCooldown":0,"spCost":0,"bb":{"hp_ratio":0.02},"bbStr":{"branch_id":"boss_summon_enemy","enemy_key":"enemy_1005_yokai"}}	branch_id → waves[t].branches
notCountInTotal, tags	false, ["sarkaz"]
be, beFactor, attrPower, isFlyEnemy	6620, 1, 6620, false	attrPower = the official float32 attribute power P = atk·5 + maxHp + def·3 + res·3 of the enemy_database record (season override NOT applied: 灼藤 / 元核孽生者 count with their database ATK) — the input of the replacement count (§10); beFactor = randomEnemyAttributeDict.enemyBattleEffectivenessFactor; be = round(P'/f) on the spawned stats (informative); isFlyEnemy = the official movement class
tokenOnly	false	only ever spawned by other enemies (summons / blackboards / level refs) — a token for 联防 routing (research 08 §5)
acTypes, acType	[], null	特训 type(s): FLY/TIMES/ELEMENT/DOT/INVISIBLE/REFLECTION/SPECIAL
templateSlot	"N"	placeholder slot when the key is a template key
summons[], inactiveIn[], seasonOverride		spawned enemies; modes banning it; overridden fields
iconId, spine	"enemy_1422_lrsldr"	asset ids

10. factions.json — special enemies (特训敌人)

Key	Example	Meaning
templateSlots	{"N":"enemy_1422_lrsldr","E":"enemy_1427_lrnazg","S":"enemy_1425_lrcmra","NF":"enemy_1005_yokai","EF":"enemy_1042_frostd","SF":"enemy_1040_bombd","T":"enemy_1000_gopro_2","TF":"enemy_1041_lazerd"}	placeholder keys in wave templates
types[type]	{"type":"FLY","name":"特训敌人·飞行","desc":"…","icon":"fly_icon","sortId":2,"typeIdentifier":2,"involveRandom":true,"count":3,"weight":3,"pool":[…],"entries":[…]}	7 types
entries[key]	{"key":"enemy_1041_lazerd","type":"FLY","weight":10,"firstHalf":true,"fly":true,"N":[{"key":"enemy_10083_hlbird"}],"E":[{"key":"enemy_1005_yokai_2"}],"inactiveIn":[]}	67 entries (specialEnemyInfoDict); fly = the movement class shared by the special, normal and elite keys; inactiveIn = modes whose inactiveEnemyKey bans the SPECIAL key (attached keys are never filtered)
generation	{"specialEnemyNum":3,"maxLevelCnt":15,"fillType":"SPECIAL","typeSlots":{"FLY":3,…},"firstHalfMaxRound":7,"minReplacedEnemyCount":1,"maxReplacedEnemyCount":5,"minActionIntervalRatio":0.05,"beFactors":{…},"placeholders":{"enemy_1422_lrsldr":{"slot":"N","cls":"normal","fly":false},…},"powerFormula":"…","countFormula":"…","timingFormula":"…","notes":[…]}	parameters of the official generator (client RandomEnemyGenerater, research 08 §2 — normative)

The official generator (implemented by server/match/waves.js):

1. Match: 3 random types (shuffle the 6 involveRandom types, keep 3); each owns typeSlots[t] (3) of the 15 round slots, fillType (SPECIAL) the other 6; the slots are shuffled → one type per round. All 15 round picks are drawn at match start (weighted entry of that type and half whose SPECIAL key the mode allows; random attached N / E).
2. Round: every template SPAWN action whose key is a placeholders key becomes the pick's normal / elite / special key, or is skipped when that key's movement class differs from the placeholder's (ground rounds drop NF/EF/SF, FLY rounds drop N/E/S). Its count n' = clamp(roundHalfEven(f32(f32(f32(n·P(tpl))/f(tpl)) / f32(P(new)/f(new)))), 1, 5) (P = enemies[k].attrPower, f = beFactor), unit i at time + i·max(n·interval/n', 0.05·n·interval). Literal keys and the T / TF token keys are kept.
3. Leader / hidden rounds use the pick of their own slot (14 / 15; solo 标准 9): only E or EF escorts spawn.

11. waves.json — { [templateId]: Wave } (38 templates)

Non-boss rounds use act1autochess_01…07, h01…h06; leaders h07_0X (+_s solo; boss_5 is act2autochess_h07_05); hidden core h08_0X; 联防 act1autochess_escaped_single|multi; training tr01/02/04.

Field	Example	Meaning
id, kind, solo	"act1autochess_03", "normal", false	kind normal / boss / hidden / escaped / training
bossId	null / "boss_8"
maxPlayTime	55	real seconds of the forced-2× battle (DESIGN §4; spawn times below are game seconds)
dp, characterLimit, moveMultiplier, bgm	{"init":10,"perSec":1,"max":99}, 8, 0.5, "bat_kazimierz2_2"	level options
routes[]	{"motion":"FLY","start":[9,10],"end":[9,2],"checkpoints":[[9,9],[12,9],…]}	WALK routes have 0–1 checkpoints (usually the end) ⇒ pathfind between them on the stage grid; motion:"E_NUM" (start/end [0,0]) = dummy route of a non-SPAWN action (only ACTIVATE_PREDEFINED in boss_4 branches)
routes[].steps	[{"t":"move","p":[1,17]},{"t":"disappear"},{"t":"wait","s":3},{"t":"appear","p":[5,10]}]	present only when a route has non-MOVE checkpoints (move, patrol, wait s, disappear, appear p); authoritative when present
routes[].spawnRandom	[0.3,0.3]	random spawn offset (tiles), optional
extraRoutes[]	same shape	routes used by branches
spawns[]	{"time":6,"key":"enemy_1422_lrsldr","count":10,"interval":6,"routeIndex":0,"slot":"N"}	time = absolute game seconds of the first unit; unit i spawns at time + i·interval. Optional: slot (placeholder N/E/S/NF/EF/SF/T/TF → replace per factions), tag (boss / part), hidden/hiddenGroup, group/pack/weight (random spawn group: group:"local" = only on the local side), unharmful (counts as killed), action (non-SPAWN action, e.g. ACTIVATE_PREDEFINED with key = device alias)
branches[name]	{"boss_summon_enemy":[[{"time":0,"key":"enemy_1005_yokai","count":1,"interval":1,"routeIndex":5}]]}	script-triggered spawns (skills[].bbStr.branch_id, dragon = 炎佑, boss_escaped_walk/fly, boss_battle_multi_player = mirrored boss sharing HP); array of phases, times relative to the phase start, routeIndex → extraRoutes. [ASSUMED] phase i runs on the i-th trigger (clamped to the last)
overrides[enemyKey]	{"enemy_1430_lrrook":{"stats":{"maxHp":60000}}}	per-template enemy overrides (bosses): partial stats, talents, skills
devices[]	{"key":"trap_039_dstnta","alias":"trap_039_dstnta#1","pos":[5,9],"dir":"…","hidden":true}	template predefines (activated by branches)
totalCount, slotCounts	22, {"N":10,"E":1,"NF":9,"EF":2}	SPAWN counts before replacement (excl. unharmful)
usedBy[]	{"modeId":"mode_multi_normal","round":3,"bossId":null}

Timing: templates have one wave with one fragment, so time = wave.preDelay + fragment.preDelay + action.preDelay exactly (the builder warns if a multi-fragment template ever appears).

12. stages.json — { [stageId]: Stage } (11 terrains, 8 active)

Field	Example	Meaning
id, name	"act2autochess_m01", "战场#05(下半) 源石流发生装置"	player-facing name from research 05 (the official tables carry none; falls back to id). The build drops bracketed segments containing Latin letters (research notes such as 战场#01 (upper half #01) → 战场#01), logs a warning, and fails validation if Latin text remains
weight, active, modes	50, true, ["mode_single_normal",…]	match-start pick weight (act1 m05–m07 weight 0)
size	[19,21]
rows[]	rows[9] = "##Errr#rrrSrrr#rrrS##"	19 strings, index = row (0 = bottom), one glyph per col
tiles[glyph]	{"tileKey":"tile_road","height":"LOW","buildable":"ALL","passable":"ALL","groundPassable":true,"flyPassable":true,"special":null,"bb":{}}	actual tile properties of each glyph used
devices[]	{"key":"trap_013_blower","name":"源石流发生装置","alias":"trap_013_blower#001","pos":[13,5],"dir":"DOWN","hidden":false,"role":"blower","active":true,"stats":{…},"rangeGrid":[[0,0],…],"rangeTiles":[[13,5],[12,5],[11,5],[10,5]],"skill":{"skillId":"sktok_blower","bb":{"blower_s_character[equal].atk":0.3,…}},"desc":null}	predefined devices; role ∈ crate, platform, mound, blower, mireController, tideController, turret (双眼皮, activated by band 机械援助), coldWind, sandstorm, waterPlatform, sealedFloor, bush, bossSpawn; active = present at match start (= not hidden in the level file); rangeGrid is facing right, rangeTiles = absolute tiles after rotating by dir (RIGHT (dr,dc), UP (dc,−dr), LEFT (−dr,−dc), DOWN (−dc,dr)), clipped to the grid (null without a range). Terrain cards / band effects toggle devices by alias
mapChars[]	{"key":"char_605_cmedic","alias":"char_605_cmedic#1","pos":[10,2],…}	band_amedic map characters (→ tokens.json)
special	{"mire":{"intervalSec":3,"aspdPerStack":-0.05,"moveMulPerStack":-0.05,"maxStacks":10,…}}, {"infection":{"bb":{"damage":70,"atk":0.2,"attack_speed":20,"duration":300}}}, {"deepsea":{"skillId":"sktok_tidectrl_3","bb":{…}}}, {"smog":{"rule":"operatorsNotTargetableByEnemyRanged"}}, {"blower":{…}}	special-terrain parameters
runes[], globalBuffs[]	{"key":"map_tile_blackb_assign","bb":{…},"bbStr":{…}}	level runes (act2 m03 reed rune has no reed tiles — leftover)
deployTiles	{"normal":{"melee":[[9,3],…],"rangedOnly":[[10,4],…],"changedByDevices":[…]},"bossLeft":{…},"bossRight":{…}}	deployable tiles at match start: melee = LOW buildable ALL/MELEE (ranged may use them too), rangedOnly = high ground / active platforms; tiles under active crates/mounds excluded
groundPaths / groundPathsWithDevices	{"12,10->9,2":[[12,10],[12,9],…,[12,3],[11,3],[10,3],[9,3],[9,2]]}	the ground route of the sim's own server/sim/grid.js flow field (research 08 §3.4: 4-direction SPFA from the goal, crates cost 1000, Bresenham line-of-sight smoothing — official route lengths — plus the remake's road-over-floor preference: among equal-length routes the fewest non-blockable floor / gate tiles, and smoothing never cuts across floor the grid route does not walk, e.g. 战场#01's lower gate takes the col-8 road, not the col-9 floor lane; DESIGN §14 corrections) as the tiles an enemy crosses, over the whole map, without / with the match-start devices (crates cost 1000, platforms / mounds blocking [ASSUMED]); the battle re-reads its flow fields live
options, config	{"characterLimit":8,"moveMultiplier":0.5}, {…configBlackBoard strings…}

Glyph legend (rows):

Glyph	tileKey	Height	Build	Ground	Meaning
#	tile_forbidden	HIGH	NONE	no (fly yes)	void
X	tile_forbidden	HIGH	NONE	no (fly no)	hard separator (rows 6, 13)
r	tile_road	LOW	ALL	yes	deployable ground
R	tile_road	LOW	NONE	yes	non-deployable road
f	tile_floor	LOW	NONE	yes	lane col 9 / preview pen
p	tile_floor	LOW	NONE	yes	preview row (previewNotAlloed)
h	tile_wall	HIGH	RANGED	no	high ground
b	tile_fence_bound	LOW	ALL	no	fenced lowland (deployable, blocks ground)
a	tile_achand	HIGH	ALL	no	hand slot (row 7 cols 0–9; row 0 boss benches)
A	tile_achand	HIGH	ALL	no	temp hand slot (row 8 cols 4–8)
S / E	tile_start / tile_end	LOW	NONE	yes	red gate / blue objective
I / O	tile_telin / tile_telout	LOW	NONE	yes	teleport in / out (boss field)
m g d i	tile_mire / tile_smog / tile_deepsea / tile_infection	LOW	ALL	yes	special terrain (see special)

13. bosses.json — { [bossId]: Boss } (10; boss_8–10 hidden)

Field	Example (boss_5)	Meaning
bossId, enemyKey, handbookId, name, sortId	"boss_5", "enemy_2016_csphtm", …, "卢西恩，“猩红血钻”", 5
weight, hidden	10, false	pick weight within its round
bloodPoint	{"FUNNY":200000,"NORMAL":390000,"HARD":780000,"ABYSS":3000000}	shared leader HP pool per difficulty
lpr	30	LP cost if it leaks
templates[modeId]	{"round":14,"template":"act2autochess_h07_05"}
escortsByTemplate[templateId]	[{"key":"enemy_2009_csaudc","slot":null,"count":4},…]	non-boss spawns (slot = still a placeholder)
parts[]	["enemy_9014_acstma","enemy_9015_acstmb"]	boss parts (random local groups / unharmful)
abilities[]	handbook texts

14. tokens.json — { [tokenId]: Token } (22)

kind: summon (19 chess summons), bondSummon (enemy_9012_acloon 炎佑 for 炎 6/9), mapChar (char_605_cmedic 预备干员-医疗 / char_613_acmedc Touch placed by band band_amedic).

Field	Example (token_10028_vigil_wolf)	Meaning
tokenId, kind, name, appellation, desc, descRaw	…, "summon", "狼群"
profession, subProfessionId, position	"TOKEN", "notchar1", "MELEE"
displayType, placeable	"DEFAULT", true	shopStateTokenDict: DEFAULT = hand-placeable piece; HIDDEN = battle-only
owners[]	["chess_char_3_19_a","chess_char_3_19_b"]
stats, rangeGrid, dmgType, attackKind, projectile, canHitFly	first owner's values	defaults
skill	{"skillId":"sktok_vigil_wolf_3","bb":{…}}	default token skill (same slot as the owner's skill)
deployLimit, count	1, 1	count = copies sent to the hand / spawned (talent/skill cnt); null ⇒ use deployLimit
variants[ownerChessId]	{"phase":2,"level":1,"stats":{…},"immunities":{…},"rangeGrid":…,"trait":{…},"dmgType":…,"skill":{full skill record},"talents":[…],"count":1,"sources":["talent","display"]}	stats at the owner's phase/level (clamped to the token's max level) + golden module tokenAttributeBlackboard; the owner's module parts flagged isToken upgrade the variant's trait (+moduleDesc) and talents (伺夜's wolves, 缪尔赛思's 流形 scale 1, 浊心斯卡蒂's 海嗣 30 s, “耀阳” atk_scale 1.15). sources ⊆ talent/skill/display: how the owner produces it (display only = listed on the character but unused by its default skill/talents, e.g. 迷迭香 S2, 凛御银灰 eagle1/3). count = copies from a talent cnt or the default skill's cnt when that skill overrides this token; null ⇒ use deployLimit
variants[o].bySkill[i]	{"skill":{…},"count":1,"sources":["talent","display"]}	owner loadout with the non-default skill index i (one entry per other selectable owner skill): the token skill of that slot (伺夜's wolves, 缪尔赛思's 流形, 凛御银灰's eagles…), the count and how the chess then produces it (sources may be []: 风丸 S1 makes no 纸偶; 赫默 / 巫恋 S1 only display ⇒ no hand piece). The sim resolves them for an owner loadout: simdata getToken(id, ownerChessId, loadout) → def.sources / def.count
variants[o].byModule[m]	{"stats":{…},"immunities":{…},"trait":{…},"talents":[…]}	golden owner with another module m or 'none': the token as that module makes it (module tokenAttributeBlackboard, isToken trait/talent parts)
assets	{"avatar":"token_10028_vigil_wolf","spine":"token_10028_vigil_wolf"}
mapChar extras	phase, level, trait, talents, positions:[{"alias":"char_613_acmedc#2_multi_only","pos":[10,10],"dir":"RIGHT","multiOnly":true}], source
bondSummon extras	bondId, motion, skills, talents, assets.isEnemyModel	炎佑 base template; ATK/HP are replaced at battle start by 30 % of the 炎 operators' sums (bonds.yanShip)

---

15. Anomalies found while joining (also in .cache/build-data-report.json)

1. 凛御银灰 container token token_10057_svash2_eagle (talent tokenKey) is not in character_table: the talent's tokenKey is remapped to the default skill's token (S2 → …_eagle2, containerTokenKey keeps the official id). …_eagle1/3 stay listed (sources:["display"], belong to S1/S3). With another skill selected (DESIGN §16) the talent follows that skill's eagle (simdata loadoutRecord; variants[o].bySkill[i].sources).
2. DIY chess (chess_char_5_diy1/2, chess_char_6_diy1/2, _a and _b = 8 records): no charId, no stats/skill; visible:false, name placeholder 甄选干员. Out of scope for v1.
3. Module-less chess: 蒂比 (chess_char_2_13) and 凛御银灰 (chess_char_5_14) have no module; their golden record has module:{id:null,active:false} and no module stat bonus.
4. Hidden chess (17) are kept with visible:false; several operators exist in two tiers with one hidden (锡人, 耶拉, 录武官, 白面鸮, 百炼嘉维尔, 魔王, 华法琳, 妮芙).
5. Every non-DIY chess (258 records: 129 normal + 129 golden) has a resolvable default skill, stats and range — no skill anomalies.
6. Skill triggers: official rows with skillIndex 0 apply to skill 1 only (research 03 C1): only 古米 and 灰毫 get TAKE_DAMAGE; bearer/tactician/bard/librator/phalanx rows match no pool chess.
7. Trait candidate rangeId (送葬人, 松果 1-3; 风丸, 归溟幽灵鲨 x-4) is the trait-effect area, exposed as trait.rangeGrid, not the attack range.
8. Passive skills use numeric spType 8 in skill_table → normalized to ON_DEPLOY.
9. Bond effectDesc keeps positional placeholders by design (effectDescParams).
10. Hidden module talents: module data-only talent candidates have no name/text; kept as hidden:true,fromModule:true.
11. act1 m02 crates and platforms are all hidden in the 下半 level file and absent from the PRTS 下半 screenshot ⇒ no crates at start (active = !hidden for every device; research 08 §3.3 — an earlier build kept #001–#004 active, which blocked the top road). Its terrain cards can only act on hidden devices. Platforms/mounds block ground movement [ASSUMED]; crates are obstacle-like (cost 1000), not walls.
12. Boss per-template overrides (e.g. 愧悔魂灵圣杯 HP 60 000, 昆图斯 ATK 480, 卢西恩 ATK 700/600) live in waves[t].overrides, not in enemies.json.
13. act1autochess_h08_01 totalCount 15 excludes the two unharmful hand parts (research counted 17).
14. Escaped (联防) templates have maxPlayTime 1 in data; use the round's combatTimeLimit.
15. 炎佑 name in the enemy DB is "炎佑" with literal ASCII quotes.
16. Research-only fields (null without docs/research): stage name, item category/kind/family/implFormula/ requiresBondId/rangeGrid/flavor, bond spec, E2 art availability (falls back to "char has an E2 phase").
17. [ASSUMED] content (flagged in data): 机变 family schedule and server pools, the solo leader pool factor (bossHpScale.solo 0.25), title criteria, income cap 12, per-turn band-draft timer 12 s. The special-enemy generator, the co-op leader pool (bloodPoint, no alive-player scaling) and the 联防 timing are official.
18. Module parts flagged isToken (伺夜, 浊心斯卡蒂, 缪尔赛思, 耀骑士临光 golden) upgrade the summon only; they are applied to tokens.json variants, never to the operator's talents/trait.
19. Undefined enemy-database fields (m_defined:false): a zero m_value means "never set" and falls back to the default (lifePointReduce 1 — 萨卡兹王庭军战士, 深池逐火战士, 圣堂剑士…; attackSpeed 100 — 卢西恩, 萨卡兹枯朽战车…; baseAttackTime 1 — 枯朽之种); a non-zero m_value is the prefab value and is kept (boss-part immunities). Defined values are never changed: 假想敌：铳/管/弦 have an official aspd 0 (skill-driven; engines clamp).
20. MELEE enemies with an official radius (粉碎攻坚手 2.5, 宿主士兵 2.5, 迷路的巨像 2.5, 清明/堂皇 2, 雪怪小队破冰者 1.9, 冰爆源石虫 1.75, 深池方阵步兵 1.5, 深池伙友卫队 1.4, 山海众头目 1.2 …): rangeRadius 0, value in rawRangeRadius.
21. Token counts: a skill cnt is only used as a token count when that skill overrides the token (夕's S1 cnt 2 is a charge count ⇒ “小自在” count null).
22. 联防 spawn timing (client GenerateHelpBattleEnemyData → _GetSpEnemyActionData → _CalculateActionPredelayConsiderUid, decoded from GameAssembly.dll; server/match/waves.js buildUniteWave): every leaked enemy is a 1-unit action on the route of the escaped template's first lrsldr (walker) / yokai (flyer) / gopro_2 / lazerd (token) action, else action 0. Per host action, with W = count·interval of that action and M = the largest per-owner group, step = min(max(W/M, 0.05·W), 5 s); the k-th owner (first-leak order) starts 0.5·k s after the action's preDelay (no cap) and its i-th unit spawns i·step later.
23. Preview zones (waves.js gateOf, client AutoChessEnemyPreviewManager): target row = route start row + 6 (leader / hidden rounds + 13); the upper pen zone (anchor (18,7)) needs target row ≥ 18, everything else goes to the lower zone (15,7). Normal fields: only the (12,10) gate is upper; boss fields: (5,10) upper, the (2,10) gate and the leaders starting at (3,x) / (4,x) lower.

16. Counts (current build)

chess 266 (112 visible; 283 selectable skills over the visible chess, 184 module choices over 129 goldens), bonds 23, garrisons 249 (43 effect keys), items 115, bands 40, effects 361, enemies 249, factions 67 entries, waves 38, stages 11 (8 active), bosses 10, tokens 22, choice events 109, bounty cards 129, tactic cards 43.

17. Integrity guarantees (checked by the builder and test/data.test.js)

Chess bonds/garrisons/tokens/base/golden ids resolve; talent tokens are in chess.tokens; every non-DIY chess has skills[] with exactly one default equal to skill (and every skill token listed in chess.tokens); golden module choices are consistent (one default iff module.active, base fields present; composing the default reproduces the record); token variants carry bySkill / byModule for exactly the other owner choices; bond members exist and carry the bond; items/bands/bonds reference existing effects; wave spawn keys and branch keys resolve in enemies.json, route indices exist; every mode round has a template (or boss templates) present in waves.json; stages are 19 × 21 with known glyphs and contiguous helper paths; bosses/factions/tokens/choice pools resolve; token variants carry sources; enemies have maxHp/bat > 0, aspd/moveSpeed/lpr ≥ 0 and rangeRadius ≥ 0 (0 for MELEE); no non-finite numbers; total size < 6 MB. test/data.test.js additionally re-derives every chess/enemy stat from the raw official tables (incl. the float32 attrPower) and rebuilds offline to prove data/ is not stale (both skipped without the cache); faction entries carry no k copies and share one movement class; every device's active = !hidden; the helper lanes of act1 m02 / m04 / act2 m02 are the official ones (research 08 §3.2). test/match/waves-official.test.js checks the generator against 429 official entry × round compositions (test/fixtures/official-waves.json); with the cache, test/match/waves-crosscheck.test.js re-implements the client generator and preview zones on the raw level files for every mode × round × template (leader and hidden rounds included) × allowed entry, and test/sim/pathing-crosscheck.test.js re-implements the client SPFA + smoothing over the whole map for every walkable tile of the normal / 联防 / boss fields of the 8 stages (extended by the road-over-floor preference; against the pure official algorithm: identical route lengths, never more non-blockable tiles crossed).