A specialized investigation skill for high-stakes analysis of mainland Chinese and Hong Kong entities with strict guards against data fabrication and character-level typos.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install cn-client-investigation
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install cn-client-investigation using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is CN Client Investigation?
The CN Client Investigation skill is a rigorous framework designed for professional analysts and bankers targeting the Chinese market, including A-share, STAR Market, ChiNext, and HK-listed companies. It addresses critical AI limitations such as token-drift typos in Chinese characters and hallucinated financial data by enforcing a multi-layered verification protocol. This skill is essential for generating investor-ready deliverables like MD reports and PPTX decks where character-level accuracy and data provenance are non-negotiable.
By prioritizing UTF-8 literals over Unicode escapes and utilizing a pre-computed Chinese lexicon, the skill bypasses common LLM encoding errors. It establishes a clear data hierarchy, fetching information from authoritative sources like CNINFO, Tushare Pro, and Tianyancha while mandating a 'provenance gate' that blocks the delivery of any report containing unverified figures.
CN Client Investigation Use Cases
- Conducting deep-dive investigations into A-share, STAR Market, or private Chinese unicorns.
- Generating banker-grade pitch decks or investment memos for Hong Kong and Mainland China targets.
- Auditing financial data across multiple Chinese regulatory and market data providers.
- Preventing 'token-drift' typos in technical Chinese financial terminology during automated document generation.
How CN Client Investigation Works
- Scope & Lexicon Initialization: Confirms the target entity type (e.g., A-share) and loads a specialized Chinese lexicon for industry-specific terms to ensure text integrity.
- Tiered Data Collection: Sequentially queries T1 (CNINFO/Tushare), T2 (SSE/SZSE/HKEX), and T3 (Wind/Tianyancha) sources to gather financial and corporate data.
- Cross-Check & Snapshotting: Verifies every hard number against at least two independent sources and saves raw MCP JSON snapshots in a local audit trail.
- Analysis Synthesis: Generates technical analysis focused on China-specific dimensions like shareholding structures (state-owned vs. private) and regulatory risks.
- Automated QA Gates: Runs mandatory Python scripts to scan for Chinese character typos and ensures every digit in the final report is linked to an entry in the data-provenance.md file.
- Compilation: Uses pptxgenjs to compile final slide decks, triggering a final post-write gate that must pass before the file is marked as shippable.
CN Client Investigation Setup
Install the skill into your environment and ensure the following MCP dependencies are active:
# Install required MCP tools
openclaw install aigroup-market-mcp aigroup-tianyancha-mcp
# Configure Tushare Pro API key for T1 data access
export TUSHARE_TOKEN='your_token_here'
# Run the provenance verification gate manually if needed
python3 skills/cn-client-investigation/scripts/provenance_verify.py deliverable/analysis.md deliverable/data-provenance.md
CN Client Investigation Data Schema & Taxonomy
The skill organizes deliverables into a strict directory structure to ensure an audit trail:
| Directory/File | Description |
|---|---|
raw-data/ |
Contains raw JSON snapshots from MCP calls (e.g., {ts_code}-market-mcp.json). |
analysis.md |
The core Markdown report containing synthesized insights and hard data. |
data-provenance.md |
A tracking table linking every number to its source, URL, and retrieval time. |
slides/ |
JavaScript source files for each slide using UTF-8 literals for Chinese text. |
references/ |
Includes cn-lexicon.js and standard data source templates. |
CN Client Investigation Advanced Features
- Estimate-as-T1 Smuggling Detection: A strict mode check that prevents agents from labeling derived estimates as primary source data.
- Precision Drift Warning: Detects and warns when the same data point appears with inconsistent rounding across different sections of a report.
- Zero-Escape Enforcement: A compilation gate that fails if any Chinese text is encoded as Unicode escape sequences instead of UTF-8 literals.
- Banker-Memo Dispatcher: A prompt-driven framework that generates 8-section research memos and content-driven slide outlines automatically.
SKILL.md
Loading
Related Openclaw Skills
AI Image to Video — Convert Images Into Video Clips
A professional AI-powered animation tool that converts static images into high-quality 1080p video clips via a cloud-based rendering pipeline.
AgentYield
AgentYield is a high-performance observability layer that detects waste and optimizes costs for LLM and tool calls without compromising data privacy.
Match Loop
Match Loop is an advanced iterative orchestration pattern for Openclaw Skills that pairs a generator agent with a visual analyst to ensure code matches target UI and functional requirements perfectly.
HD Video Maker Free Download
A high-performance AI video editing skill that transforms raw clips into polished 1080p HD videos through simple text commands and cloud-based rendering.
Awesome Design Skill
A professional UI design style selector that integrates over 50+ industry-leading design systems as development guides for AI coding agents.
AgentPay SDK
A self-custodial SDK and CLI for automating crypto transfers, merchant payments via Bitrefill, and policy-driven wallet management.
Featured*
