简历标准化解析虾

简历标准化解析虾 — 将非标简历（PDF/Word/图片/文本）解析为结构化候选人数据，并可写入飞书多维表格或导出 Excel/JSON。 **当以下情况时使用此 Skill**： (1) 用户上传或提供简历文件（PDF、Word、图片），要求提取候选人信息 (2) 需要批量解析简历并录入数据库或飞书多维表格 (...

Security Scan

The skill's code and instructions match a resume-parsing utility and there are no signs of hidden exfiltration or unrelated privileges, though it references Feishu integrations and requires local Python libraries that you should verify before use.

Purpose & Capability

The name/description, included resume-schema and skills dictionary, and the parse-resume.py script all align: extracting text from PDF/DOCX/images, parsing fields, scoring quality, and exporting to JSON/Excel or writing to a table. No unrelated services or credentials are requested in the package itself.

Instruction Scope

SKILL.md stays within parsing, normalization, scoring, and output workflows. It references calling feishu_im_bot_image to download files and feishu_bitable_app_table_record to write to Feishu tables and mentions triggering other '虾' (skills). Those references imply integration with external platform connectors but the instructions do not attempt to read arbitrary system paths or exfiltrate data to unknown endpoints.

Install Mechanism

There is no install spec (instruction-only skill) and only a bundled Python script. The script relies on common Python packages (pdfplumber, python-docx, pytesseract/Pillow, pandas/openpyxl) and will exit with a message if they are missing. No downloads from unknown URLs or archive extraction are present.

Credentials

The registry metadata declares no required env vars or credentials, but SKILL.md references writing to Feishu (and using Feishu-specific helper names). That implies the agent/platform or another skill must provide Feishu credentials; the skill itself does not declare or request them. Also the tool processes PII (names, phones, emails); treat credential and data access accordingly.

Persistence & Privilege

always is false and the skill does not request persistent system-wide privileges or modify other skills' configurations. It may invoke other skills (e.g., interview-scheduler-claw) which is normal collaboration behavior.

Assessment

This package appears coherent for resume parsing, but before installing/checking it you should: (1) confirm how Feishu integration is authorized on your platform (are credentials supplied by the agent runtime or will you need to provide tokens?), (2) be aware the tool handles PII (phones/emails) — ensure storage, retention, and access controls meet your privacy policies, (3) install and verify required Python dependencies (pdfplumber, python-docx, pytesseract + tesseract OCR engine, Pillow, pandas, openpyxl) and OCR language data, (4) test on non-sensitive sample resumes to validate parsing and deduplication behavior, and (5) review how/where original files are stored or uploaded (e.g., if the platform automatically uploads files to external services). If you need higher assurance, request explicit documentation of Feishu connectors or an implementation that declares required credentials and data flow.

将非标简历转化为标准化候选人数据，支持单份和批量处理，可输出到飞书多维表格、Excel 或 JSON。

工作流程

[简历文件/文本] → [格式识别] → [内容提取] → [字段解析] → [标准化] → [输出/入库]

步骤 1：获取简历内容

输入类型	处理方式
PDF（文本层）	用 scripts/parse-resume.py 提取文本
PDF（扫描件）/ 图片	先 OCR，再提取文本
Word/DOCX	用 scripts/parse-resume.py 解析段落和表格
纯文本	直接进入字段解析
飞书消息中的文件	用 feishu_im_bot_image 下载到本地，再解析

步骤 2：字段解析

解析目标字段（参考 references/resume-schema.md）：

• 基本信息：姓名、手机、邮箱、性别、年龄、所在城市
• 求职意向：期望职位、期望薪资、到岗时间
• 教育经历：学校、专业、学历、时间
• 工作经历：公司、职位、时间段、职责描述（自动计算工作年限）
• 技能标签：技术技能、语言能力、证书（参考 references/skills-dictionary.md 标准化）
• 项目经历：项目名称、角色、技术栈、成果

步骤 3：标准化处理

• 时间格式统一为 YYYY-MM
• 学历统一为枚举：本科 / 硕士 / 博士 / 专科 / 高中
• 技能名称标准化（如 "JS" → "JavaScript"，参考 references/skills-dictionary.md）
• 工作年限自动计算（当前年份 - 最早工作年份）
• 对置信度低的字段标记 [待确认]

步骤 4：质量评分

• 完整度评分（0-100）：必填字段覆盖率
• 逻辑校验：工作时间是否重叠、学历时间是否合理
• 核心字段（姓名、手机/邮箱）缺失时标记为"低质量简历"

步骤 5：输出

根据用户需求选择输出方式：

输出方式	操作
飞书多维表格	用 feishu_bitable_app_table_record 写入，表结构参考 references/resume-schema.md
Excel	用 scripts/parse-resume.py export 导出
JSON	直接输出标准化 JSON
飞书消息摘要	格式化后直接回复用户

批量处理

用户提供多份简历时：

1. 逐份解析，汇总结果
2. 相同手机号/邮箱的候选人自动去重（保留最新版本）
3. 最终输出汇总表格，注明解析成功/失败数量

与其他虾的协作

• 解析完成后，可直接触发 面试邀约虾（interview-scheduler-claw）发送面试邀请
• 标准化数据可传给 简历筛选虾 按条件筛选候选人

参考文件

• references/resume-schema.md — 标准字段定义和飞书多维表格建表结构
• references/skills-dictionary.md — 技能名称标准化词典
• scripts/parse-resume.py — 文件解析脚本（PDF/Word/图片）