wangweih

credit-information-query

CreditChina enterprise credit information report download, Chinese enterprise dishonesty/risk-list checks, and bidding/procurement evidence preservation for Codex agents. Use when the user needs to query an explicit company on CreditChina, China Enforcement Information Disclosure Network, or China Government Procurement Network; download official credit information report PDFs; convert PDFs to JPG; or preserve webpage/PDF/JPG evidence with normalized output names. 中文补充:用于信用中国信用信息查询、法院失信被执行人、中国政府采购网严重违法失信记录和招投标信用截图留痕。

wangweih 0 Updated 3w ago

Resources

4
GitHub

Install

npx skillscat add wangweih/credit-information-query

Install via the SkillsCat registry.

SKILL.md

Credit Information Query / 信用信息查询

Purpose

Use this skill to preserve enterprise credit evidence from Chinese public websites:

  • CreditChina (creditchina.gov.cn)
  • China Enforcement Information Disclosure Network (zxgk.court.gov.cn)
  • China Government Procurement Network (ccgp.gov.cn)

The bundled script opens the target site, searches for the explicit enterprise name, handles captcha workflows where possible, captures trace screenshots/HTML, exports PDF evidence, converts PDF pages to JPG, and normalizes final bid_review evidence names.

中文补充:本 Skill 自动完成企业信用查询、网页截图、PDF 导出、JPG 转换和招投标证据留存。

Non-Negotiable Invocation Rules

  • Always provide an explicit enterprise name.
  • Always provide an explicit query type.
  • Do not assume a default company.
  • Do not assume a default function.
  • First infer the user intent, then choose one query type.

中文补充:脚本没有默认企业,也没有默认功能。Agent 必须先判断需求,再显式选择查询类型。

Quick Start

python scripts/creditchina_query.py "企业名称" 查询类型

Check dependencies:

python scripts/creditchina_query.py --check-env

Install dependencies:

pip install -r scripts/requirements.txt

List query types:

python scripts/creditchina_query.py --list-types

Query Type Selection

Choose exactly one:

  • 信用信息查询 / credit_report: download the official CreditChina credit information report PDF and convert it to JPG pages.
  • tax: query CreditChina major tax violation dishonesty subjects.
  • govproc: query CreditChina government procurement serious violation dishonesty list.
  • court_shixin: query dishonesty judgment debtor information on zxgk.court.gov.cn; preserve the CreditChina court-jump trace first.
  • ccgp_cr: query government procurement serious violation dishonesty records on ccgp.gov.cn.
  • bid_review: run tax, govproc, court_shixin, and ccgp_cr; then create normalized bidding-review evidence in separated pdf/ and jpg/ folders.

中文补充:招投标/比选申请人信用查询通常使用 bid_review;单独下载信用报告使用 信用信息查询

Common Commands

Credit information report:

python scripts/creditchina_query.py "杭州申卓农业开发有限公司" 信用信息查询 --no-manual

Bidding/procurement evidence retention:

python scripts/creditchina_query.py "杭州申卓农业开发有限公司" bid_review --no-manual

Custom output directory:

python scripts/creditchina_query.py "杭州申卓农业开发有限公司" bid_review --output-dir=output

Output Contract

Default output root:

scripts/output/

Relative --output-dir values are resolved from the script directory.

The script separates final evidence:

  • PDFs go under pdf/.
  • JPGs go under jpg/.
  • The manifest stays in the output root.
  • HTML backups, process screenshots, and captcha debug images remain as trace/supporting files.

For bid_review, final evidence must follow this naming style:

00_查询证据留存清单_杭州申卓农业开发有限公司.txt
jpg/01_信用中国_重大税收违法失信主体_杭州申卓农业开发有限公司.jpg
jpg/02_信用中国_政府采购严重违法失信行为记录名单_杭州申卓农业开发有限公司.jpg
jpg/03_信用中国_失信被执行人跳转提示_杭州申卓农业开发有限公司.jpg
jpg/04_中国执行信息公开网_失信被执行人_第1页_杭州申卓农业开发有限公司.jpg
jpg/05_中国执行信息公开网_失信被执行人_第2页_杭州申卓农业开发有限公司.jpg
jpg/06_中国政府采购网_政府采购严重违法失信行为记录名单_杭州申卓农业开发有限公司.jpg

PDF files use matching normalized names under pdf/.

中文补充:最终交付重点是 pdf/jpg/00_查询证据留存清单_企业名称.txt

Agent Workflow

  1. Parse the user request for the exact enterprise name.
  2. Decide whether the task is:
    • a CreditChina credit information report download,
    • a single list query,
    • or a full bidding/procurement evidence package.
  3. Select the explicit query type.
  4. Run --check-env if this is a new machine/session.
  5. Run the script with explicit arguments.
  6. Verify that expected PDF/JPG files exist.
  7. Report the output directory and key evidence files.

中文补充:Agent 完成后应核对 PDF/JPG 是否生成,再报告路径。

Operational Notes

  • Keep the browser window visible; minimized windows can break screenshots.
  • Use --no-manual for unattended runs. Omit it when manual captcha fallback is allowed.
  • No-record query results are still valid evidence if screenshots/PDF/JPG are produced.
  • Do not create ZIP archives unless the user explicitly asks for one.
  • If Python resolves to a Windows Store placeholder, run the script with the actual runtime Python executable and re-check dependencies.
  • If the CreditChina court-jump prompt detection changes, preserve the current CreditChina special-query page as the jump trace and continue to the court site instead of failing the entire bid_review.

中文补充:验证码可能需要重试;网站结构变化时,保留当前页面证据并尽量继续后续流程。