Receipt Assistant

报销票据处理助手。自动识别票据（火车票、打车发票、行程单、酒店发票）、提取关键信息（日期、金额、发票号）、按规则重命名文件、生成汇总报表。触发词：报销、发票、票据识别、火车票、打车票、酒店发票、整理票据。

Security Scan

The skill's code, instructions, and required packages align with its stated purpose (local receipt scanning, OCR via the platform image tool, filename generation, and Excel report output); it does not request unrelated credentials or network endpoints.

Purpose & Capability

Name/description (receipt processing: detect, extract, rename, report) matches the included Python modules (scanner, namer, reporter, utils). Declared requirements are minimal (no credentials, no external services). The SKILL.md instructs using the platform 'image' tool for visual recognition — consistent with an OCR workflow.

Instruction Scope

SKILL.md and code limit actions to scanning a user-provided directory, calling the platform image tool for OCR, extracting fields, renaming/ copying files, and writing an Excel report. There are no instructions to read unrelated system files or to transmit data to external endpoints outside the expected image tool and local filesystem.

Install Mechanism

Registry has no automated install spec (instruction-only), but the package includes requirements.txt and SKILL.md suggests running `pip install openpyxl pdfplumber Pillow`. These are standard PyPI packages appropriate for PDF/image parsing and Excel generation. No downloads from untrusted URLs or archive extraction are present.

Credentials

The skill does not require credentials. It will optionally read COMPANY_NAME, PERSON_NAME, and OUTPUT_DIR environment variables (and a local config.json) — all directly relevant to naming/output behavior. No sensitive tokens or unrelated env vars are requested.

Persistence & Privilege

always is false and the skill does not attempt to modify other skills or global agent settings. It writes local configuration (config.json) and output files under the project/output directory by default — expected for a file-processing tool.

Assessment

This skill appears coherent and limited to local receipt processing, but consider these precautions before installing: 1) The SKILL.md asks you to run `pip install` for standard libraries (openpyxl, pdfplumber, Pillow) — install in a virtualenv to avoid polluting your system Python. 2) The skill reads/writes files: it will create/modify config.json and copy/rename files into its output directory (default under the project). Point it at a test folder first and back up originals. 3) OCR is done via the platform 'image' tool per the instructions — be aware images/receipt data will be processed by that tool (check privacy/policy for the platform). 4) Inspect config.json and the included code if you want to confirm there are no additional behaviors in your deployment. 5) No network endpoints, credentials, or unexpected installers are present, so running in a restricted environment or VM is a reasonable precaution but not strictly required.

报销助手

自动处理报销票据：识别、提取信息、重命名、生成报表。

工作流程

1. 扫描目录 → 找到票据文件（PDF/JPG/PNG/JPEG）
2. 视觉识别 → 调用 image 工具识别票据
3. 提取信息 → 解析日期、金额、发票号、销售方等
4. 重命名文件 → 按规则生成标准化文件名
5. 生成报表 → 输出 Excel 汇总表

首次配置

引导用户配置：

• 公司名称（用于发票命名）
• 默认姓名（用于火车票命名）
• 输出目录

票据命名规则

类型	格式
火车票	{姓名}-{日期}-{金额}-中国铁路
打车发票	{公司}-{日期}-{发票号}-{金额}-{销售方}-发票
行程单	{公司}-{日期}-{发票号}-{金额}-{销售方}-行程单
酒店发票	{公司}-{日期}-{发票号}-{金额}-{销售方}

视觉识别 Prompt

火车票

请识别这张火车票的以下信息，以 JSON 格式返回：
{
  "type": "火车票",
  "date": "乘车日期（YYYYMMDD）",
  "amount": "票价金额（只数字）",
  "person": "乘客姓名"
}
只返回 JSON。

打车发票/行程单

请识别这张票据的以下信息，以 JSON 格式返回：
{
  "type": "打车发票 或 行程单",
  "date": "开票日期（YYYYMMDD）",
  "invoice_no": "发票号码",
  "amount": "金额（只数字）",
  "seller": "销售方名称"
}
只返回 JSON。

酒店发票

请识别这张酒店发票的以下信息，以 JSON 格式返回：
{
  "type": "酒店发票",
  "date": "开票日期（YYYYMMDD）",
  "invoice_no": "发票号码",
  "amount": "金额（只数字）",
  "seller": "酒店名称"
}
只返回 JSON。

输出结构

output/
└── YYYYMMDD_HHMMSS/
    ├── 报销汇总.xlsx
    ├── 张三-20240327-533-中国铁路.pdf
    └── ...

脚本接口

函数	用途
scan_directory(path)	扫描目录，返回文件列表
generate_filename(...)	生成标准化文件名
generate_excel_report(tickets, output_dir)	生成 Excel 汇总
check_config()	检查配置状态
update_config(...)	更新配置

安装依赖

pip install openpyxl pdfplumber Pillow