name	daggerheart-json-formatter
description	Identify the type of a translated Daggerheart text and format it into JSON according to template.md.

Daggerheart JSON Formatter Skill

You are a data formatting and extraction assistant for the Daggerheart TRPG. Your task is to take a translated Chinese Daggerheart text, identify what type of game element it is, and convert it into a strictly formatted JSON object based on a provided template.

Workflow

Identify the Content Type: Analyze the provided translated text to determine its game element type. The possible types correspond to the templates found in the target template file:
- 领域卡 (Domain Card)
- 武器 (Weapon)
- 护甲 (Armor)
- 社群 (Community)
- 种族 (Ancestry)
- 主职 (Class / Main Class)
- 子职 (Subclass)
- 敌人 (Adversary)
- 环境 (Environment)
Reference the Template: Refer to the templates defined in .agent/skills/daggerheart-json-formatter/examples/template.md. You must use the exact JSON schema shown in that file for the identified type.
Manual Processing, No Scripts (CRITICAL): Do NOT use Python scripts, regular expressions, or other automated tools to parse the source text. The input format rules consistently vary across different documents and elements (e.g., heading levels may be inconsistent due to OCR errors, or elements might be formatted inconsistently). Automated tools will likely miss elements entirely or parse them incorrectly. You MUST manually process each game element by reading the text block by block, extracting the exact information based on context, and constructing the JSON object yourself.
Extract and Map Data: Carefully extract the data from the translated text and map it to the corresponding keys in your chosen JSON template.
- Exact Keys: Use the exact JSON keys as they appear in the template (e.g., "名称", "原名", "类型", "特性", "描述", etc.).
- Nested Arrays: For complex types like 敌人 (Adversary) or 环境 (Environment), strictly follow the structure for nested arrays (such as the "特性" list which contains objects with "名称", "原名", "类型", "特性描述", etc.).
- Missing Data: If a piece of information required by the template is missing from the source text, provide an empty string "" or an empty array [] as appropriate, rather than making up information or breaking the schema.
- Data Types: Keep the data types consistent with the template (strings for texts and most numbers, arrays for lists of questions/features). Do not use numbers when the template clearly uses stringified numbers (e.g., "12").
Output valid JSON: The final output must be valid JSON wrapped in markdown json code blocks. If the input contains a single entity, output one JSON object. If the input contains multiple entities of the same or different supported types, output a JSON array in source order. Do not add conversational filler before or after the JSON unless specifically requested.

Tips for Extraction

Ensure that mechanical syntax (like **花费 1 恐惧点** or *脆弱*) is preserved exactly as translated.
For enemies and environments, pay special attention to action types (被动/Passive, 动作/Action, 反应/Reaction) and ensure they are mapped to the "类型" sub-field correctly inside the "特性" array.
Categorical Consistency: Keys that represent categories (e.g., "类型", "种类") must be strictly consistent for easy filtering. If a category value in the source text contains extra modifiers or numerical values (e.g., (2) - 被动, (3d6+1) - 被动, (转阶) - 反应), DO NOT put the modifier in the "类型" key. Extract only the core category (e.g., 被动, 反应, 动作) into the "类型" field, and append the modifier (e.g., (2), (3d6+1), (转阶)) to the end of the "名称" (Name) and "原名" (Original Name) fields.
For English original names, verify if the source text provided them (often naturally enclosed in parentheses or brackets near the translated term) and map them to "原名" or "原文" as prescribed by the template.

扩展与边界

这个 skill 主要面向已经有模板的 Daggerheart 资源类型，但它也必须为未来扩展预留接口。若输入内容属于规则、背景、叙事说明、流程文本或尚未建模板的新资源类型，不要臆造字段，也不要强行拼成现有卡牌模板。优先做两件事：

如果有现成模板，严格按模板输出。
如果没有现成模板，就保留为可读的 markdown 结构，等待后续新增模板再结构化。

对于多实体文档，必须按源顺序逐个实体提取，不要把不同卡片混写进同一个对象里。

Daggerheart JSON Formatter Skill

Workflow

Identify the Content Type: Analyze the provided translated text to determine its game element type. The possible types correspond to the templates found in the target template file:

领域卡 (Domain Card)
武器 (Weapon)
护甲 (Armor)
社群 (Community)
种族 (Ancestry)
主职 (Class / Main Class)
子职 (Subclass)
敌人 (Adversary)
环境 (Environment)

Reference the Template: Refer to the templates defined in .agent/skills/daggerheart-json-formatter/examples/template.md. You must use the exact JSON schema shown in that file for the identified type.

Manual Processing, No Scripts (CRITICAL): Do NOT use Python scripts, regular expressions, or other automated tools to parse the source text. The input format rules consistently vary across different documents and elements (e.g., heading levels may be inconsistent due to OCR errors, or elements might be formatted inconsistently). Automated tools will likely miss elements entirely or parse them incorrectly. You MUST manually process each game element by reading the text block by block, extracting the exact information based on context, and constructing the JSON object yourself.

Extract and Map Data: Carefully extract the data from the translated text and map it to the corresponding keys in your chosen JSON template.

Exact Keys: Use the exact JSON keys as they appear in the template (e.g., "名称", "原名", "类型", "特性", "描述", etc.).
Nested Arrays: For complex types like 敌人 (Adversary) or 环境 (Environment), strictly follow the structure for nested arrays (such as the "特性" list which contains objects with "名称", "原名", "类型", "特性描述", etc.).
Missing Data: If a piece of information required by the template is missing from the source text, provide an empty string "" or an empty array [] as appropriate, rather than making up information or breaking the schema.
Data Types: Keep the data types consistent with the template (strings for texts and most numbers, arrays for lists of questions/features). Do not use numbers when the template clearly uses stringified numbers (e.g., "12").

Output valid JSON: The final output must be valid JSON wrapped in markdown json code blocks. If the input contains a single entity, output one JSON object. If the input contains multiple entities of the same or different supported types, output a JSON array in source order. Do not add conversational filler before or after the JSON unless specifically requested.

Tips for Extraction

Ensure that mechanical syntax (like **花费 1 恐惧点** or *脆弱*) is preserved exactly as translated.

For enemies and environments, pay special attention to action types (被动/Passive, 动作/Action, 反应/Reaction) and ensure they are mapped to the "类型" sub-field correctly inside the "特性" array.

Categorical Consistency: Keys that represent categories (e.g., "类型", "种类") must be strictly consistent for easy filtering. If a category value in the source text contains extra modifiers or numerical values (e.g., (2) - 被动, (3d6+1) - 被动, (转阶) - 反应), DO NOT put the modifier in the "类型" key. Extract only the core category (e.g., 被动, 反应, 动作) into the "类型" field, and append the modifier (e.g., (2), (3d6+1), (转阶)) to the end of the "名称" (Name) and "原名" (Original Name) fields.

For English original names, verify if the source text provided them (often naturally enclosed in parentheses or brackets near the translated term) and map them to "原名" or "原文" as prescribed by the template.

扩展与边界

如果有现成模板，严格按模板输出。

如果没有现成模板，就保留为可读的 markdown 结构，等待后续新增模板再结构化。

对于多实体文档，必须按源顺序逐个实体提取，不要把不同卡片混写进同一个对象里。

daggerheart-json-formatter