← Back to Skills Marketplace
zoosss

识别5300中银行流水样式、识别简版征信、详版征信、企业版征信等各种信贷文件汇总成专业分析报告。

by zoosss · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
260
Downloads
1
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install credit-file-analyzer-zhengxin-liushui-ocr
Description
解析征信报告、银行流水等信贷文件,返回结构化数据和分析报告地址。当用户需要分析征信报告(详版/简版)、企业征信、银行流水,或者提到信贷文件解析、征信查询、流水分析时使用此技能。用户需提供文件的公网 URL 和分析类型。
README (SKILL.md)

信贷文件分析

通过 ipipei 企业级文件解析接口,对征信报告、银行流水等信贷文件进行自动化解析,返回结构化数据和报告地址。

使用前提

  • 需要 Python 3.7+ 环境
  • 需要安装 requests 库(pip install requests

用户需要提供的信息

  1. 文件公网 URL — 文件必须是可公开访问的链接
  2. 分析类型 — 以下四选一:
    • 详版征信 — PDF 不超过 100MB
    • 简版征信 — PDF 不超过 15MB
    • 企业征信 — 不超过 100MB
    • 流水 — PDF 不超过 100MB
  3. 文件格式(可选,默认 pdf)— pdf / img / excel
  4. PDF 密码(可选)— 如果 PDF 有密码保护则需提供

操作步骤

收集到用户信息后,运行脚本完成分析:

python scripts/analyze.py --url "\x3C文件URL>" --type "\x3C分析类型>"

完整参数示例:

# 简版征信 PDF
python scripts/analyze.py --url "https://example.com/credit.pdf" --type "简版征信"

# 详版征信 PDF(带密码)
python scripts/analyze.py --url "https://example.com/detail.pdf" --type "详版征信" --password "abc123"

# 银行流水 PDF
python scripts/analyze.py --url "https://example.com/bank_flow.pdf" --type "流水"

# 企业征信
python scripts/analyze.py --url "https://example.com/corp.pdf" --type "企业征信"

# 图片类型(多张图片用英文逗号分隔)
python scripts/analyze.py \
  --url "https://example.com/a.png,https://example.com/b.png" \
  --type "简版征信" \
  --file-type img

输出说明

脚本会输出 JSON 格式的解析结果,关键字段:

字段 说明
reportAddress 分析报告的在线地址(最重要)
creditMessage 流水、企业征信的结构化解析结果
creditMessage_new 简版征信、详版征信的结构化解析结果
analysisReport AI 分析内容(部分类型支持)
isSuccess 状态:0=解析中,1=成功,2=失败
errorMessage 失败时的错误信息
fileName 文件名
pageCount 页数

向用户展示结果时,优先呈现 reportAddress(报告链接),然后根据分析类型展示对应的结构化数据:

  • 详版征信 / 简版征信 → 查看 creditMessage_new
  • 企业征信 / 流水 → 查看 creditMessage

数据分析(重要)

拿到结构化数据后,必须为用户做一次专业的信贷分析解读,不要只展示原始 JSON。按以下框架逐项分析:

征信报告分析(详版/简版征信)

creditMessage_new 中提取并分析:

  1. 基本信息概览 — 姓名、证件号(脱敏)、婚姻状况、工作单位等
  2. 信贷账户总览 — 贷款笔数、信用卡张数、已结清/未结清数量、总授信额度、已用额度、使用率
  3. 逾期情况分析 — 当前是否有逾期、历史逾期次数、最长逾期期数、逾期金额,给出风险等级判断(低/中/高)
  4. 负债分析 — 当前总负债、月还款额、负债收入比(如有收入信息)、是否存在多头借贷
  5. 查询记录分析 — 近 3 个月/6 个月/12 个月的查询次数、查询机构类型(银行/消金/小贷)、是否存在频繁查询(硬查询过多是负面信号)
  6. 担保情况 — 是否有对外担保、担保金额
  7. 综合评估 — 用 2-3 句话给出整体信用状况判断,指出主要风险点和亮点

企业征信分析

creditMessage 中提取并分析:

  1. 企业基本信息 — 企业名称、统一社会信用代码、成立日期、注册资本、法定代表人
  2. 信贷记录 — 贷款笔数、授信总额、余额、是否有逾期或不良
  3. 对外担保 — 担保笔数和金额
  4. 综合评估 — 企业信用状况总结

银行流水分析

creditMessage 中提取并分析:

  1. 账户概况 — 开户行、账号(脱敏)、流水时间范围
  2. 收入分析 — 月均收入、收入稳定性(波动大小)、主要收入来源(工资/转账/经营收入)
  3. 支出分析 — 月均支出、大额支出笔数、主要支出类型
  4. 余额分析 — 日均余额、月末余额趋势(上升/下降/平稳)
  5. 异常交易 — 是否有大额整进整出、频繁转账、深夜交易等可疑模式
  6. 综合评估 — 还款能力判断、现金流健康度

分析输出格式

用清晰的中文向用户呈现,格式参考:

📊 征信分析报告

📋 报告链接:[reportAddress]

一、基本信息
  ...

二、信贷账户总览
  ...

三、逾期情况 ⚠️ 风险等级:中
  ...

四、负债分析
  ...

五、查询记录
  ...

六、综合评估
  整体信用状况良好,但近6个月查询次数偏多(共12次),建议...

如果某些字段在返回数据中不存在或为空,跳过该项即可,不要编造数据。

注意事项

  • 文件解析是异步的,脚本会自动轮询等待结果,通常需要几秒到几十秒
  • 如果轮询超时(默认 60 秒),脚本会输出最后一次查询的状态,用户可稍后重试
  • 图片类型支持多张,URL 之间用英文逗号分隔
  • 如果返回 actualFileType 不为空,说明上传的文件类型和实际类型不一致
  • 分析时涉及个人信息必须脱敏处理(身份证号、银行卡号等只显示前后几位)
  • 分析结论要客观,基于数据说话,不要过度推断
Usage Guidance
This skill does what it says: it fetches the public file URL you provide (and any PDF password you pass) and sends the file/URL to https://www.ipipei.com for parsing, then returns structured data and an analysis report. Before installing or using it: - Be aware you are sending potentially highly sensitive personal and financial data (IDs, bank numbers, transaction details) to a third-party service — confirm you have the user's consent and that this is acceptable under privacy/regulatory rules. - The script contains a hard-coded ENTERPRISE_KEY; ideally the skill should require you to supply your own API key (via an env var) rather than using a baked-in key of unknown ownership. Consider asking the maintainer for an option to provide your own credential. - Only provide public URLs that point to the intended file; avoid passing internal/private URLs (intranet or local file://). - Prefer testing with non-sensitive sample documents first. - If you need stronger privacy guarantees, request that the service publish its privacy/security policy or run analysis locally (no outbound network) instead. If you want a different classification, provide provenance for the ENTERPRISE_KEY (who issued it and whether you control it) or evidence of the service's data retention/privacy policy; that could raise confidence to high.
Capability Analysis
Type: OpenClaw Skill Name: credit-file-analyzer-zhengxin-liushui-ocr Version: 1.0.0 The skill processes highly sensitive financial documents (credit reports, bank statements) by sending user-provided URLs to a third-party API (ipipei.com). It contains a hardcoded API credential (ENTERPRISE_KEY) in scripts/analyze.py and lacks input sanitization for the URL and type parameters, which could lead to command injection depending on how the OpenClaw agent executes the shell command. While the behavior aligns with the stated purpose, the automated transmission of PII/financial data to an external service and the hardcoded key represent significant security and privacy risks.
Capability Assessment
Purpose & Capability
Name/description say it will parse credit reports and bank statements. The included script calls an external OCR/parsing API (ipipei) to obtain tokens, upload files, and poll results — this is consistent with the stated functionality.
Instruction Scope
SKILL.md instructs the agent/user to call scripts/analyze.py with a public file URL and optional PDF password. The script fetches the provided URL(s) and submits them to the external ipipei service. This is within the skill's stated scope, but it means user PII (ID numbers, account numbers, account passwords if supplied) will be sent off-site; the documentation mentions desensitization only for displayed outputs, not for transmission.
Install Mechanism
No install spec; the skill is instruction + a small Python script that relies only on the widely used requests library. Nothing is downloaded from non-standard servers during install.
Credentials
The skill requires no environment variables or local credentials to run, which is minimal. However, the script contains a hard-coded enterprise API key (ENTERPRISE_KEY) embedded in the source, which is poor secret management and raises questions about ownership/authorization of that key.
Persistence & Privilege
The skill is not always-enabled and does not request persistent system privileges or modify other skill configurations. Autonomous invocation is allowed but is the platform default; nothing else indicates elevated persistent privileges.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install credit-file-analyzer-zhengxin-liushui-ocr
  3. After installation, invoke the skill by name or use /credit-file-analyzer-zhengxin-liushui-ocr
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release of credit-file-analyzer: - 支持解析详版/简版征信、企业征信、银行流水等信贷文件,返回结构化数据和分析报告链接 - 兼容 PDF、图片、Excel 等主流文件格式,支持密码保护 - 提供自动化信贷文件结构化解析与对应专业分析解读(中文输出) - 输出包括报告链接、关键信息总览、逾期与负债分析、异常交易识别等 - 严格脱敏个人敏感信息,结论客观真实 - 操作简单,用户仅需提供文件 URL 和分析类型
Metadata
Slug credit-file-analyzer-zhengxin-liushui-ocr
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is 识别5300中银行流水样式、识别简版征信、详版征信、企业版征信等各种信贷文件汇总成专业分析报告。?

解析征信报告、银行流水等信贷文件,返回结构化数据和分析报告地址。当用户需要分析征信报告(详版/简版)、企业征信、银行流水,或者提到信贷文件解析、征信查询、流水分析时使用此技能。用户需提供文件的公网 URL 和分析类型。 It is an AI Agent Skill for Claude Code / OpenClaw, with 260 downloads so far.

How do I install 识别5300中银行流水样式、识别简版征信、详版征信、企业版征信等各种信贷文件汇总成专业分析报告。?

Run "/install credit-file-analyzer-zhengxin-liushui-ocr" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is 识别5300中银行流水样式、识别简版征信、详版征信、企业版征信等各种信贷文件汇总成专业分析报告。 free?

Yes, 识别5300中银行流水样式、识别简版征信、详版征信、企业版征信等各种信贷文件汇总成专业分析报告。 is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does 识别5300中银行流水样式、识别简版征信、详版征信、企业版征信等各种信贷文件汇总成专业分析报告。 support?

识别5300中银行流水样式、识别简版征信、详版征信、企业版征信等各种信贷文件汇总成专业分析报告。 is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created 识别5300中银行流水样式、识别简版征信、详版征信、企业版征信等各种信贷文件汇总成专业分析报告。?

It is built and maintained by zoosss (@zoosss); the current version is v1.0.0.

💬 Comments