Cn Ib Cross Document Check logo

Cn Ib Cross Document Check

Community
QiushenZou
cn-ib-cross-document-check

Reconcile facts, figures, dates, names, definitions, and disclosures across multiple Chinese investment-banking documents. Use when comparing 招股说明书、发行保荐书、上市保荐书、审计报告、法律意见书、募集资金运用文件、问询回复、再融资申报材料 or different document versions for 跨文件一致性、数据勾稽、口径核对、版本差异 or contradiction detection. Use cn-ib-filing-proofreader for single-document language and formatting review.

Overview

PublisherQiushenZou
Repositorycn-investment-banking-skills
Skill namecn-ib-cross-document-check
Stars
97
Forks
0
Bundled files
5
LicenseApache-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 5 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by QiushenZou on GitHub. Read the source before you install it.

Installation

Install the Cn Ib Cross Document Check AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/QiushenZou/cn-investment-banking-skills/tree/main/skills/cn-ib-cross-document-check
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/QiushenZou/cn-investment-banking-skills.git /tmp/cn-investment-banking-skills
mkdir -p .claude/skills
cp -r /tmp/cn-investment-banking-skills/skills/cn-ib-cross-document-check .claude/skills/cn-ib-cross-document-check
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Cn Ib Cross Document Check in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Cn Ib Cross Document Check on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Cn Ib Cross Document Check is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

CN IB Cross-Document Check

Compare multiple filing documents without flattening differences in period, entity, scope, unit, or accounting definition.

Non-negotiable rules

  1. Compare like with like. Match entity, metric definition, period, scope, currency, unit, and pre/post-transaction basis before declaring a conflict.
  2. Preserve provenance for every extracted fact: document, version, page/location, source text, and extraction method.
  3. Treat rounding differences separately from substantive inconsistencies.
  4. Do not decide which document is correct merely because one is newer. Identify the likely source of truth and request confirmation.
  5. Distinguish a missing disclosure from a contradictory disclosure.
  6. Escalate legal, accounting, and regulatory judgments to the responsible professional.

Workflow

1. Build the document manifest

Read references/document-matrix.md. Record for every file:

  • document type;
  • filename and version;
  • document date;
  • reporting period;
  • responsible owner;
  • expected source-of-truth fields;
  • whether the document is draft, signed, filed, feedback response, or registration version.

Never compare files until their versions are identified.

2. Define the comparison dictionary

Start with the high-risk fields relevant to the transaction:

  • issuer and subsidiary names;
  • incorporation and historical dates;
  • share capital, ownership, and per-share figures;
  • revenue, profit, cash flow, assets, liabilities, and margins;
  • customers, suppliers, related parties, and contracts;
  • market share, ranking, capacity, production, and sales volume;
  • fundraising project investment, proceeds, implementation entity, and schedule;
  • risk factors and material event descriptions;
  • sponsor, counsel, accountant, and signatory details.

Define each metric before extraction. Read references/comparison-schema.md.

3. Extract an evidence ledger

Create one row per claim using:

text
document, version, entity, metric, definition, period, scope,
value, unit, location, source_text, extraction_method, reviewer

Do not normalize away the original text. Store both the original value/unit and the normalized value/unit.

4. Run deterministic comparison

For structured numeric claims, use:

bash
python3 scripts/compare_numeric_claims.py claims.csv --output conflicts.csv

The script normalizes common Chinese currency and share units and applies a configurable tolerance. It does not determine the authoritative value.

5. Review contextual conflicts

Read references/conflict-rules.md. For every script result or manually observed difference, decide whether it is:

  • a true contradiction;
  • a period mismatch;
  • an entity or consolidation-scope mismatch;
  • a pre/post-transaction mismatch;
  • a currency or unit mismatch;
  • a rounding difference;
  • a definition mismatch;
  • an intentional update;
  • a missing disclosure;
  • insufficient evidence.

Quote the minimum relevant source text from each document.

6. Identify the likely source of truth

Use the document matrix and transaction context. Examples:

  • audited financial figures usually trace to the latest signed audit report and underlying ledger;
  • share capital traces to current corporate records and the applicable transaction assumption;
  • fundraising amounts trace to the approved project and filing version;
  • legal entity names trace to valid registration documents;
  • market data traces to the cited external source and specified date.

Phrase the conclusion as 建议以……为准,并由……确认, not as an unsupported final determination.

7. Produce the reconciliation report

Use this structure:

IDSeverityTopicDocument ADocument BDifferenceClassificationLikely source of truthRequired actionOwner

Also provide:

  • counts by severity and classification;
  • unresolved P0/P1 items;
  • documents or appendices not received;
  • fields excluded from comparison and why;
  • a closure column for corrected version and reviewer evidence.

8. Apply approved corrections

When asked to correct existing DOCX files:

  • preserve each original and edit the identified authoritative version only;
  • apply native Word tracked insertions and deletions by default;
  • change the smallest necessary text, number, unit, or table cell;
  • preserve existing revisions and comments unless instructed otherwise;
  • do not simulate redlines with color or strikethrough;
  • record the corrected version and exact location in the reconciliation report;
  • structurally verify the revision markup, re-render changed pages, and rerun the comparison.

Use a transparent revision author such as Codex when none is specified. Produce a clean accepted copy only when explicitly requested.

Completion gate

Do not call the review complete until:

  • every document has a recorded version;
  • all P0/P1 items have an owner and action;
  • every numerical conflict retains source locations;
  • rounding and definition differences are distinguished;
  • corrections are rechecked in the revised files;
  • visual or scanned content excluded from extraction is disclosed.

Use temporary storage for extraction and comparison files. Keep only user-requested final artifacts in the deal directory.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Cn Ib Cross Document Check AI skill do?

Reconcile facts, figures, dates, names, definitions, and disclosures across multiple Chinese investment-banking documents. Use when comparing 招股说明书、发行保荐书、上市保荐书、审计报告、法律意见书、募集资金运用文件、问询回复、再融资申报材料 or different document versions for 跨文件一致性、数据勾稽、口径核对、版本差异 or contradiction detection. Use cn-ib-filing-proofreader for single-document language and formatting review.

Why use Cn Ib Cross Document Check on TypingMind?

Because you install it once and use it with any model. Cn Ib Cross Document Check is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Cn Ib Cross Document Check in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/QiushenZou/cn-investment-banking-skills/tree/main/skills/cn-ib-cross-document-check. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Cn Ib Cross Document Check?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Cn Ib Cross Document Check?

As many as you like. As long as a model supports skills, you can use Cn Ib Cross Document Check with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Cn Ib Cross Document Check AI skill free?

Yes. It is published on GitHub by QiushenZou under the Apache-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇