中考短文填空识别与格式化

中考短文填空识别与格式化技能。当用户上传中考短文填空题目图片（和答案图片）并要求识别、格式化、输出Word或推送飞书时触发。适用场景：(1) 上传题目图片+答案图片 (2) 要求识别并格式化短文填空 (3) 要求输出Word文档（答案置于题目下方）(4) 要求存入飞书多维表格（答案置于右侧列）(5) 要求推送至飞书

Security Scan

The skill's instructions, scope, and requirements are internally consistent for an OCR → format → export workflow; it is an instruction-only skill that relies on platform-provided tools for image understanding and messaging and does not request unexplained credentials or installs.

Purpose & Capability

Name/description (OCR short-cloze recognition, format, export to DOCX/Feishu) align with the SKILL.md steps: use an image-understanding tool, normalize text and candidate words, build a DOCX and/or create/push a Feishu multidimensional table. There are no declared environment variables or unrelated binaries requested, which matches an instruction-only skill that expects platform tools to provide OCR/messaging.

Instruction Scope

Instructions stay focused on OCR, formatting and exporting. They reference two abstract platform tools: images_understand (OCR) and message (send DOCX / table link). The SKILL.md does not instruct reading arbitrary local files or unrelated environment variables. Note: the skill assumes the agent/platform supplies the OCR and Feishu messaging capabilities; it does not specify authentication/permission flows for pushing to Feishu or where generated DOCX is stored.

Install Mechanism

No install spec and no code files are present. This is the lowest-risk form: nothing is downloaded or written by an installer step.

Credentials

The skill declares no environment variables or credentials. That is proportionate for an instruction-only skill that uses platform-provided tools. If the platform's Feishu/message tool requires credentials, those are expected to be managed by the platform rather than by the skill itself.

Persistence & Privilege

always is false and the skill does not request persistent system-level privileges or modification of other skills. Autonomous invocation is allowed (default) but that is normal and not by itself a concern.

Assessment

This skill appears coherent and is instruction-only: it will OCR uploaded images, reformat text and candidate words, produce a DOCX, and (if requested) create and push a Feishu table/link using platform tools. Before installing or invoking: 1) Confirm how your platform handles Feishu authentication and whether you must grant the agent permission to create/send documents — the SKILL.md assumes a messaging tool but does not declare credentials. 2) Be aware you are sending uploaded images (possible exam content/answers) to the platform's OCR and messaging tools; avoid uploading sensitive personal data. 3) Test using non-sensitive sample images so you can verify the formatting, storage location of generated DOCX, and the exact permissions that will be used to create/push Feishu tables. 4) If you require on-premise or private Feishu credentials, ask how those credentials are stored/used by the agent (the skill itself does not request them). If anything about the Feishu push or storage locations is unclear, ask the skill author or platform admin for details before use.

核心流程

题目图片 ──OCR识别──▶ 标准化题目
答案图片 ──OCR识别──▶ 答案列表
         ↓
    按渠道输出
    ├── Word：答案置于题目下方
    └── 飞书多维表格：答案置于右侧列

执行步骤

Step 1：识别题目图片

使用 images_understand 工具：

prompt: 请准确识别这张图片中所有文字，保持原有段落结构、数字序号（36. 37. 38. 等）、下划线、空格和标点符号。特别注意：题目编号、方框内的备选词、原文段落都要完整保留。

Step 2：识别答案图片（如有）

若用户同时上传了答案图片，立即用同一工具识别：

prompt: 请准确识别这张图片中的所有答案内容，提取每个填空编号对应的答案单词，例如：36.find, 37.camera...。如果答案图片中包含题目正文中的单词，请逐个列出编号和对应答案。

Step 3：格式还原规则

备选词格式转换：

• 原始（空格分隔）：by camera clearly delicious find future great offer rush way
• 目标（顿号分隔）：by、camera、clearly、delicious、find、future、great、offer、rush、way

段落结构：

• 保留原文所有换行、段落结构
• 保留题目编号：36.～45.
• 保留所有下划线 ____
• 保留原文标点、大小写、括号注释

Step 4A：输出 Word（答案置于题目下方）

创建 DOCX，结构如下：

【标题】中考短文填空-标准化题目

【备选词】
by、camera、clearly、delicious、find、future、great、offer、rush、way

【短文正文】
[还原后正文]

【答案】（如有）
36.find  37.camera  38.________  39.________  40.________
41.________  42.________  43.________  44.________  45.________

Step 4B：输出飞书多维表格（答案置于右侧列）

新建表格，字段顺序：

字段名	类型	说明
科目	文本
短文正文	文本
备选词	文本
填空编号	文本
答案	文本	右侧列，存放每题答案
文档链接	URL	Word文档链接

答案格式： 36.find 37.camera 38.________ ...

Step 5：推送飞书

• 文档推送：用 message 工具发送 DOCX 文件到用户飞书
• 表格推送：发送多维表格链接

注意事项

• OCR 题目时务必保持原文段落格式，不要合并段落
• 备选词统一转换为中文顿号（、）分隔
• 下划线数量必须与原图一致
• 括号注释（如 canteen(食堂)）必须完整保留
• 答案优于填空位置：若图片中同时有答案，优先提取答案单词

触发示例

• "识别这张短文填空图片，答案在第二张"
• "上传了题目和答案，推送到飞书多维表格"
• "帮我识别并输出Word，答案在图片里"