cortico-world-simple-trpg-check 0.2.0 → 0.2.2
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/LICENSE +1 -1
- package/README.md +110 -114
- package/package.json +4 -3
- package/src/ENV_PROMPT.md +3 -1
- package/src/model.ts +17 -15
- package/src/request.ts +42 -19
package/LICENSE
CHANGED
package/README.md
CHANGED
|
@@ -1,114 +1,110 @@
|
|
|
1
|
-
<!-- Owner: src/definition.ts, src/world.ts, src/config.ts, src/request.ts, src/
|
|
2
|
-
|
|
3
|
-
#
|
|
4
|
-
|
|
5
|
-
|
|
6
|
-
|
|
7
|
-
|
|
8
|
-
|
|
9
|
-
|
|
10
|
-
|
|
11
|
-
|
|
12
|
-
|
|
13
|
-
|
|
14
|
-
-
|
|
15
|
-
|
|
16
|
-
|
|
17
|
-
|
|
18
|
-
|
|
19
|
-
|
|
20
|
-
|
|
21
|
-
|
|
22
|
-
|
|
23
|
-
|
|
24
|
-
|
|
25
|
-
|
|
26
|
-
|
|
27
|
-
|
|
28
|
-
|
|
29
|
-
|
|
30
|
-
|
|
31
|
-
|
|
32
|
-
|
|
33
|
-
|
|
34
|
-
|
|
35
|
-
|
|
36
|
-
|
|
37
|
-
|
|
38
|
-
|
|
39
|
-
|
|
40
|
-
|
|
41
|
-
|
|
42
|
-
|
|
43
|
-
|
|
44
|
-
|
|
45
|
-
|
|
46
|
-
|
|
47
|
-
|
|
48
|
-
|
|
49
|
-
|
|
50
|
-
|
|
51
|
-
|
|
52
|
-
|
|
53
|
-
|
|
54
|
-
|
|
55
|
-
|
|
56
|
-
|
|
57
|
-
|
|
58
|
-
|
|
59
|
-
|
|
60
|
-
|
|
61
|
-
|
|
62
|
-
|
|
63
|
-
|
|
64
|
-
|
|
65
|
-
|
|
66
|
-
|
|
67
|
-
|
|
68
|
-
|
|
69
|
-
|
|
70
|
-
|
|
71
|
-
|
|
72
|
-
|
|
73
|
-
|
|
74
|
-
|
|
75
|
-
|
|
76
|
-
|
|
77
|
-
|
|
78
|
-
|
|
79
|
-
|
|
80
|
-
|
|
81
|
-
|
|
82
|
-
|
|
83
|
-
|
|
84
|
-
|
|
85
|
-
|
|
86
|
-
|
|
87
|
-
|
|
88
|
-
|
|
89
|
-
|
|
90
|
-
|
|
91
|
-
|
|
92
|
-
|
|
93
|
-
|
|
94
|
-
|
|
95
|
-
|
|
96
|
-
|
|
97
|
-
|
|
98
|
-
|
|
99
|
-
|
|
100
|
-
|
|
101
|
-
corepack pnpm
|
|
102
|
-
corepack pnpm typecheck
|
|
103
|
-
corepack pnpm
|
|
104
|
-
|
|
105
|
-
|
|
106
|
-
|
|
107
|
-
|
|
108
|
-
|
|
109
|
-
|
|
110
|
-
|
|
111
|
-
|
|
112
|
-
在 Cortico 仓库根目录运行 `corepack pnpm check:extension <本包绝对路径>`。安装到运行中的实例需要重启进程;本包的测试和装载检查不会启动 bot。
|
|
113
|
-
|
|
114
|
-
项目采用 MIT License,详见 [LICENSE](LICENSE)。
|
|
1
|
+
<!-- Owner: src/definition.ts, src/world.ts, src/config.ts, src/request.ts, src/model.ts, src/ENV_PROMPT.md, src/display.ts, src/contest.ts -->
|
|
2
|
+
|
|
3
|
+
# Simple TRPG Check
|
|
4
|
+
|
|
5
|
+
Cortico 的 TRPG 技能检定 World 扩展,检定规则由COC7规则启发。agent 提供角色背景、技能证据和当前情境,Laya 或 Jev 估计角色的技能水平,扩展据此进行 d100 检定,返回骰值、结果和可直接展示的文本。
|
|
6
|
+
|
|
7
|
+
支持单项检定、组合检定、对抗检定,以及奖励骰、惩罚骰和困难/极难任务。技能名可以自定义,不需要预先填写数值角色卡。
|
|
8
|
+
|
|
9
|
+
## 安装与配置
|
|
10
|
+
|
|
11
|
+
要求 Node.js 22 或更新版本,以及 Cortico World API 5。在 Cortico 仓库的 `extensions/` 目录安装:
|
|
12
|
+
|
|
13
|
+
```powershell
|
|
14
|
+
Set-Location "<Cortico 仓库路径>\extensions"
|
|
15
|
+
corepack pnpm add --ignore-workspace cortico-world-simple-trpg-check
|
|
16
|
+
```
|
|
17
|
+
|
|
18
|
+
在目标部署的“扩展”页启用本 World,重启进程,再进入 World 的配置页选择判定模型。
|
|
19
|
+
|
|
20
|
+
| 判定模型 | 配置方式 |
|
|
21
|
+
| --- | --- |
|
|
22
|
+
| Jev | 将“判定模型”设为 `jev`,选择 TypeSafe、OpenRouter 或自定义来源,填写对应密钥。自定义来源还需配置服务地址。 |
|
|
23
|
+
| 多语言 Laya(默认) | 选择已安装 `laya` 的 Conda Python 环境。首次使用可能下载模型权重,后续使用本地缓存。 |
|
|
24
|
+
| 英文 Laya | 将“判定模型”设为 `laya`,通过可选依赖 `@receptron/laya` 在本机运行。 |
|
|
25
|
+
|
|
26
|
+
现有评测中,本地 Laya 对不同技能水平证据的区分较弱,建议使用 Jev。选择后点击“测试连接”验证服务与密钥;该操作不掷骰。
|
|
27
|
+
|
|
28
|
+
Jev 密钥保存在部署目录的 `.env` 中,配置页的“打开密钥文件”可打开该文件。填写并保存后,下一次模型请求读取密钥。
|
|
29
|
+
|
|
30
|
+
| Jev 来源 | 密钥变量 |
|
|
31
|
+
| --- | --- |
|
|
32
|
+
| TypeSafe | `CORTICO_JEV_TYPESAFE_API_KEY` |
|
|
33
|
+
| OpenRouter | `CORTICO_JEV_OPENROUTER_API_KEY` |
|
|
34
|
+
| 自定义 | `CORTICO_JEV_API_KEY` |
|
|
35
|
+
|
|
36
|
+
“强制适配多语言”要求 agent 用英文填写请求内容,不翻译已有输入。切换判定模型或 Python 环境后需要重启进程。Laya 空闲释放时间默认 10 分钟;同一进程内共用模型时取各使用者中最长的保留时间。
|
|
37
|
+
|
|
38
|
+
## 使用示例
|
|
39
|
+
|
|
40
|
+
环境提示词会指导 agent 在结果不确定且情境具有戏剧性时调用检定工具。单项和组合检定使用 `simple_trpg_check_roll`:
|
|
41
|
+
|
|
42
|
+
```json
|
|
43
|
+
{
|
|
44
|
+
"character": {
|
|
45
|
+
"traits": "常年带队登山的向导。",
|
|
46
|
+
"condition": "无。"
|
|
47
|
+
},
|
|
48
|
+
"situation": {
|
|
49
|
+
"weather": "晴天,无风。",
|
|
50
|
+
"gear": "绳索与铁锁齐全,状态良好。",
|
|
51
|
+
"info": "路线与岩点位置已知。",
|
|
52
|
+
"time": "无时限。",
|
|
53
|
+
"senses": "光线充足。"
|
|
54
|
+
},
|
|
55
|
+
"checks": [
|
|
56
|
+
{
|
|
57
|
+
"skill": "攀爬",
|
|
58
|
+
"goal": "攀上岩壁,到达平台。",
|
|
59
|
+
"evidence": "靠带队攀岩谋生,处理过大量同类线路,也能应付少见的难例。"
|
|
60
|
+
}
|
|
61
|
+
]
|
|
62
|
+
}
|
|
63
|
+
```
|
|
64
|
+
|
|
65
|
+
`evidence` 描述已经确立的能力证据,例如训练经历、实操经验和同行评价,不直接填写技能数值或档位名。伤势、工具、天气等情境因素填写在对应字段,由 agent 据此选择 `bonus` 或 `grade`。`character.condition` 与 `situation` 的五个字段都必填,无影响时写“无”。
|
|
66
|
+
|
|
67
|
+
回执包含诊断结果,末尾的“显示:”块供 agent 原样展示。例如:
|
|
68
|
+
|
|
69
|
+
```text
|
|
70
|
+
【攀爬检定:80/62 失败】
|
|
71
|
+
```
|
|
72
|
+
|
|
73
|
+
两个会推理、会反应的角色互相比试时使用 `simple_trpg_check_roll_contest`。`actor` 和 `opponent` 各提供 `name`、`traits`、`condition`、`situation` 与 `checks`;每侧的 `bonus`、`joint` 填在该侧顶层。
|
|
74
|
+
|
|
75
|
+
完整工具约定见 [环境提示词](src/ENV_PROMPT.md)。
|
|
76
|
+
|
|
77
|
+
## 检定规则
|
|
78
|
+
|
|
79
|
+
模型根据 `evidence` 估计技能水平,转换成 1–99 的检定阈值,回执中称为“难度”。无修正时,骰值不超过难度即成功;不超过难度的一半为困难成功,四分之一为极难成功。1 固定为大成功,100 固定为大失败。
|
|
80
|
+
|
|
81
|
+
财富按收入、资产、生活水准和可支配资金评估,使用从贫困到富裕的档位;“信用”“信用评级”和对应英文名也可使用。
|
|
82
|
+
|
|
83
|
+
| 选项 | 行为 |
|
|
84
|
+
| --- | --- |
|
|
85
|
+
| `bonus` | 合计优势/劣势与花掉的奖励骰。通常为 −2…2;掷 `1 + |bonus|` 颗 d100,优势取最低、劣势取最高。达到 +3 自动成功,达到 −3 自动失败,均不调用模型、不掷骰。 |
|
|
86
|
+
| `grade` | 1 要求至少困难成功,2 要求至少极难成功。大成功与大失败不受影响。不能与 `joint` 同用。 |
|
|
87
|
+
| `joint` | 至少两项技能共用一次投骰。`and` 取各项难度和 `bonus` 的最小值,`or` 取最大值。 |
|
|
88
|
+
| 对抗 | 每侧投一次骰,先比成功等级,同级比难度数值,数值高者胜;再相等则 `actor` 胜。每侧多项技能必须指定该侧的 `joint`;对抗不接受 `grade`。 |
|
|
89
|
+
|
|
90
|
+
同一次普通调用中的多项技能默认独立检定。需要“先 A 再 B”时分两次调用,第二次根据变化后的事实重写情境。
|
|
91
|
+
|
|
92
|
+
奖励骰的授予和记账由 agent 通过外部工具完成,扩展不保存跨调用状态。对抗检定不触发奖励骰;某侧达到自动判定条件时,两侧都不调用模型、不掷骰,胜负规则见环境提示词。
|
|
93
|
+
|
|
94
|
+
模型不可用、答案无效或输入超出预算时,整次调用返回失败原因,不掷骰。请求按 token 估算限制长度:情境 768、单项问题 256、总输入 1000,检测时预留 10% 余量。对抗的两侧分别检测。
|
|
95
|
+
|
|
96
|
+
## 本地开发
|
|
97
|
+
|
|
98
|
+
```sh
|
|
99
|
+
corepack pnpm install
|
|
100
|
+
corepack pnpm test
|
|
101
|
+
corepack pnpm typecheck
|
|
102
|
+
corepack pnpm typecheck:web
|
|
103
|
+
corepack pnpm build:console
|
|
104
|
+
```
|
|
105
|
+
|
|
106
|
+
在 Cortico 仓库根目录运行 `corepack pnpm check:extension <本包绝对路径>` 验证扩展。构建控制台前需停止本机正在运行的 bot 进程。
|
|
107
|
+
|
|
108
|
+
评测方法、语料和离线读出训练见仓库内的 `evals/README.md`。运行时使用包内已有的读出表;`evals/` 和 `tests/` 不随 npm 包发布。
|
|
109
|
+
|
|
110
|
+
MIT License,见 [LICENSE](LICENSE)。
|
package/package.json
CHANGED
|
@@ -1,9 +1,10 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "cortico-world-simple-trpg-check",
|
|
3
|
-
"version": "0.2.
|
|
3
|
+
"version": "0.2.2",
|
|
4
4
|
"type": "module",
|
|
5
|
-
"description": "TRPG
|
|
6
|
-
"license": "MIT",
|
|
5
|
+
"description": "Cortico 的 TRPG 技能检定 World 扩展,使用 Laya 或 Jev 评估技能水平,支持 d100 掷骰、组合与对抗检定。",
|
|
6
|
+
"license": "MIT",
|
|
7
|
+
"author": "DTMosken",
|
|
7
8
|
"keywords": [
|
|
8
9
|
"cortico-world",
|
|
9
10
|
"cortico"
|
package/src/ENV_PROMPT.md
CHANGED
|
@@ -10,7 +10,9 @@
|
|
|
10
10
|
|
|
11
11
|
技能名可以用下面这套常用表(有子分类的写具体项,例如格斗的「斗殴」「刀」、射击的「手枪」「步枪」、科学的「化学」「生物学」;表外的技能名也可以用,前后保持一致):
|
|
12
12
|
|
|
13
|
-
|
|
13
|
+
力量、体质、敏捷、智力(别名灵感)、意志、教育(别名知识)、说服、取悦、恐吓、欺骗、侦查、聆听、运动、潜行(别名乔装、追踪)、巧手(别名偷窃、伪造、锁匠)、驾驶、骑乘、导航、格斗、射击、闪避、投掷、爆破、急救、医学、科学、神秘学、历史、财富(别名信用、信用评级)。
|
|
14
|
+
|
|
15
|
+
财富的 `evidence` 写已确立的收入、资产、生活水准和可支配资金,判定按贫困至富裕的档位衡量。
|
|
14
16
|
|
|
15
17
|
`situation` 的五个字段加上 `character.condition` 就是情境的全部六个维度,一个都不能省略;某项确实没有影响就写「无」。
|
|
16
18
|
|
package/src/model.ts
CHANGED
|
@@ -10,7 +10,7 @@ import {
|
|
|
10
10
|
} from './config.ts';
|
|
11
11
|
import { startMultilingualLaya } from './multilingual-laya.ts';
|
|
12
12
|
import { applyReadout, readoutFor } from './readout.ts';
|
|
13
|
-
import { LEVEL_OPTIONS, questionsFor, serializeState, type CheckRequest, type LevelQuestion } from './request.ts';
|
|
13
|
+
import { LEVEL_OPTIONS, isWealthSkill, levelBands, questionsFor, serializeState, type CheckRequest, type LevelQuestion } from './request.ts';
|
|
14
14
|
import { layaRuntimeKey, sharedLayaPool, type SharedLayaClient, type SharedLayaModel } from './shared-laya.ts';
|
|
15
15
|
|
|
16
16
|
type Questions = Record<string, LevelQuestion>;
|
|
@@ -35,17 +35,16 @@ export async function loadEnglishLaya(): Promise<SharedLayaModel> {
|
|
|
35
35
|
}
|
|
36
36
|
|
|
37
37
|
/**
|
|
38
|
-
*
|
|
39
|
-
* Jev
|
|
40
|
-
* 本地 Laya 检查点实测读不出水平证据(极端证据只挪动 15 点),保持恒等读出。
|
|
38
|
+
* 水平读出:普通技能的十档分布经读出映射为 0–100;财富的六档分布直接按代表值取期望。
|
|
39
|
+
* 普通技能的 Jev 读出使用 src/level-readout.ts,本地 Laya 保持恒等读出;运行时不训练。
|
|
41
40
|
*/
|
|
42
41
|
|
|
43
42
|
/** 读一项技能的档位概率分布并归一化;缺项按 0 计。无效时抛错。 */
|
|
44
|
-
export function probabilityVector(item: unknown, label: string): number[] {
|
|
43
|
+
export function probabilityVector(item: unknown, label: string, options: readonly string[] = LEVEL_OPTIONS): number[] {
|
|
45
44
|
const probabilities = item && typeof item === 'object' ? (item as { probabilities?: unknown }).probabilities : undefined;
|
|
46
45
|
if (!probabilities || typeof probabilities !== 'object') throw new Error(`${label}的模型答案无效`);
|
|
47
46
|
const distribution = probabilities as Record<string, unknown>;
|
|
48
|
-
const shares =
|
|
47
|
+
const shares = options.map((option) => {
|
|
49
48
|
const share = distribution[option];
|
|
50
49
|
if (share === undefined) return 0;
|
|
51
50
|
if (typeof share !== 'number' || !Number.isFinite(share) || share < 0 || share > 1) {
|
|
@@ -58,18 +57,21 @@ export function probabilityVector(item: unknown, label: string): number[] {
|
|
|
58
57
|
return shares.map((share) => share / mass);
|
|
59
58
|
}
|
|
60
59
|
|
|
61
|
-
/**
|
|
62
|
-
export function levelsFromAnswers(result: unknown,
|
|
60
|
+
/** 按技能对应的标尺读出水平,返回顺序与技能列表一致。 */
|
|
61
|
+
export function levelsFromAnswers(result: unknown, skills: readonly string[], backend: CheckBackend): number[] {
|
|
63
62
|
if (!result || typeof result !== 'object') throw new Error('模型没有返回判定结果');
|
|
64
63
|
const answers = (result as { answers?: unknown }).answers;
|
|
65
64
|
if (!answers || typeof answers !== 'object' || Array.isArray(answers)) throw new Error('模型没有返回技能答案');
|
|
66
65
|
const keyed = answers as Record<string, unknown>;
|
|
67
|
-
if (Object.keys(keyed).length !==
|
|
68
|
-
const readout = readoutFor(backend);
|
|
69
|
-
return
|
|
70
|
-
|
|
71
|
-
|
|
72
|
-
|
|
66
|
+
if (Object.keys(keyed).length !== skills.length) throw new Error('模型返回的技能答案数量不匹配');
|
|
67
|
+
const readout = readoutFor(backend);
|
|
68
|
+
return skills.map((skill, index) => {
|
|
69
|
+
const bands = levelBands(true, skill);
|
|
70
|
+
const p = probabilityVector(keyed[`check_${index}`], `第 ${index + 1} 项技能`, LEVEL_OPTIONS.slice(0, bands.length));
|
|
71
|
+
return isWealthSkill(skill)
|
|
72
|
+
? bands.reduce((sum, band, position) => sum + p[position]! * band.representative, 0)
|
|
73
|
+
: applyReadout(p, readout);
|
|
74
|
+
});
|
|
73
75
|
}
|
|
74
76
|
|
|
75
77
|
export class SystemOneSkillScorer implements SkillScorer {
|
|
@@ -84,7 +86,7 @@ export class SystemOneSkillScorer implements SkillScorer {
|
|
|
84
86
|
const result = this.ctx.cfg.backend === 'jev'
|
|
85
87
|
? await this.requestJev(state, questions)
|
|
86
88
|
: await this.getLayaClient().systemOne(state, questions);
|
|
87
|
-
return levelsFromAnswers(result, request.checks.
|
|
89
|
+
return levelsFromAnswers(result, request.checks.map((check) => check.skill), this.ctx.cfg.backend);
|
|
88
90
|
}
|
|
89
91
|
|
|
90
92
|
async close(): Promise<void> {
|
package/src/request.ts
CHANGED
|
@@ -44,11 +44,7 @@ export function checksLabel(checks: CheckSpec[], joint?: Joint): string {
|
|
|
44
44
|
}
|
|
45
45
|
|
|
46
46
|
/**
|
|
47
|
-
*
|
|
48
|
-
* 专家=硕士博士 / 世界顶尖)之间各插一档,使模型能表达两个锚点之间的位置;档位越密,
|
|
49
|
-
* 读出越不会粘在锚点上。整张表对所有判定通用,不区分技能与信用;
|
|
50
|
-
* 信用那套切点不同(0/1/10/50/90/99),这里用的是技能尺的代表值,属已知近似,
|
|
51
|
-
* 等后续用别的手段代偿情境修正时一并重估。
|
|
47
|
+
* 档位描述与代表值。普通技能按训练与专业水平分十档,财富按生活水准与资产分六档。
|
|
52
48
|
*/
|
|
53
49
|
export interface LevelBand {
|
|
54
50
|
label: string;
|
|
@@ -80,12 +76,37 @@ const LEVEL_BANDS: Record<'zh' | 'en', LevelBand[]> = {
|
|
|
80
76
|
{ label: 'One of the very few recognised as the strongest in the field', representative: 89 },
|
|
81
77
|
{ label: 'Among the best in the world at this skill', representative: 95 },
|
|
82
78
|
],
|
|
83
|
-
};
|
|
84
|
-
|
|
85
|
-
|
|
86
|
-
|
|
87
|
-
|
|
88
|
-
|
|
79
|
+
};
|
|
80
|
+
|
|
81
|
+
const WEALTH_BANDS: Record<'zh' | 'en', LevelBand[]> = {
|
|
82
|
+
zh: [
|
|
83
|
+
{ label: '身无分文,流落街头', representative: 0 },
|
|
84
|
+
{ label: '极端贫困,勉强果腹', representative: 5 },
|
|
85
|
+
{ label: '中等水平,在合理的前提下生活得比较舒适', representative: 30 },
|
|
86
|
+
{ label: '富裕,某些条件下近乎奢侈', representative: 70 },
|
|
87
|
+
{ label: '富有,生活无忧并极为奢侈', representative: 94 },
|
|
88
|
+
{ label: '超级有钱,钱对你来说只是数字', representative: 99 },
|
|
89
|
+
],
|
|
90
|
+
en: [
|
|
91
|
+
{ label: 'Penniless and living on the streets', representative: 0 },
|
|
92
|
+
{ label: 'Extremely poor, barely able to eat', representative: 5 },
|
|
93
|
+
{ label: 'Average means, living reasonably comfortably', representative: 30 },
|
|
94
|
+
{ label: 'Wealthy, with some luxuries', representative: 70 },
|
|
95
|
+
{ label: 'Rich, living carefree and very luxuriously', representative: 94 },
|
|
96
|
+
{ label: 'Extremely wealthy; money is just a number', representative: 99 },
|
|
97
|
+
],
|
|
98
|
+
};
|
|
99
|
+
|
|
100
|
+
export const LEVEL_OPTIONS = ['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J'] as const;
|
|
101
|
+
|
|
102
|
+
const WEALTH_SKILLS = new Set(['财富', '信用', '信用评级', 'wealth', 'credit', 'credit rating']);
|
|
103
|
+
|
|
104
|
+
export function isWealthSkill(skill: string): boolean {
|
|
105
|
+
return WEALTH_SKILLS.has(skill.trim().toLowerCase());
|
|
106
|
+
}
|
|
107
|
+
|
|
108
|
+
export function levelBands(chinese: boolean, skill = ''): LevelBand[] {
|
|
109
|
+
return (isWealthSkill(skill) ? WEALTH_BANDS : LEVEL_BANDS)[chinese ? 'zh' : 'en'];
|
|
89
110
|
}
|
|
90
111
|
|
|
91
112
|
export interface LevelQuestion {
|
|
@@ -163,11 +184,13 @@ export function serializeState(request: CheckRequest): string {
|
|
|
163
184
|
}
|
|
164
185
|
|
|
165
186
|
/** 每项技能一个问题:角色在这一点上处于哪一档。选项键用中性字母,档位描述进 criteria。 */
|
|
166
|
-
export function questionsFor(request: CheckRequest, chinese: boolean): Record<string, LevelQuestion> {
|
|
167
|
-
|
|
168
|
-
|
|
169
|
-
|
|
170
|
-
|
|
171
|
-
|
|
172
|
-
|
|
173
|
-
}
|
|
187
|
+
export function questionsFor(request: CheckRequest, chinese: boolean): Record<string, LevelQuestion> {
|
|
188
|
+
return Object.fromEntries(request.checks.map((check, index) => {
|
|
189
|
+
const bands = levelBands(chinese, check.skill);
|
|
190
|
+
return [`check_${index}`, {
|
|
191
|
+
type: 'choice' as const,
|
|
192
|
+
instructions: `In the skill "${check.skill}", which level is the character at? Character: ${check.evidence}`,
|
|
193
|
+
criteria: Object.fromEntries(bands.map((band, position) => [LEVEL_OPTIONS[position]!, band.label])),
|
|
194
|
+
}];
|
|
195
|
+
}));
|
|
196
|
+
}
|