jwangkun/claude-for-financial-services-cn

china-model-update

Update A-share financial models with new quarterly data, management guidance, or macro changes.

ソースを見る

使い方

  1. インストールコマンドをコピーし、自分の環境へ導入する前にソースを確認します。
  2. 必要なツール、認証情報、対応環境を原文で確認します。
  3. 入力と成果物を指定し、出力を資料と照合します。
npx skills add https://github.com/jwangkun/claude-for-financial-services-cn --skill china-model-update

開始前の確認

  • 導入にはモデル契約、API枠、ファイルへのアクセス許可は含まれません。
  • スター数や導入数は活動の指標であり、独立した品質・安全性試験ではありません。
リポジトリの原文

見出し、例、コード、表、リンク、参照画像を含む原文を表示しています。

china-model-update

Purpose

Update existing A股财务模型 with new data, ensuring all cells are traceable to sources and all changes are documented.

Data Sources

Primary: iFind MCP (Tier-1 付费) / AkShare MCP (Tier-2 免费备选)

python
get_financials(ticker, "income", "quarterly")   → Q[X] actual results
get_financials(ticker, "balance", "quarterly")  → BS update
get_financials(ticker, "cashflow", "quarterly") → CF update
get_financials(ticker, "income", "annual")      → Full year update
get_quote(ticker)                               → Current market data
get_stock_info(ticker)                          → Any company changes

Secondary Sources

  • 巨潮资讯 — official filings for exact figures
  • 业绩说明会 transcript — management commentary
  • 管理层指引 — guidance from earnings calls
  • Wind / Choice / 同花顺 — consensus updates

Workflow

Step 1: Identify What Changed

Change triggers:

  • Quarterly earnings release (季报/年报)
  • Management guidance update (管理层指引调整)
  • Macro assumption change (rate, tax, policy)
  • Model error or refinement
  • M&A or restructuring event

Change log template:

DateChange TypeItemOld ValueNew ValueReason
Earnings updateRevenue FY25EXX亿XX亿Q1 actuals beat
GuidanceTax rate25%25%No change
MacroCapEx %5%6%New plant announced

Step 2: Update Historical Actuals

Quarterly actuals:

[Company] Q[X] 20XX Actuals (from AkShare / 巨潮):
- 营业收入: XXX亿 (YoY: +XX%)
- 毛利率: XX% (vs prior: XX%)
- 归母净利润: XXX亿 (YoY: +XX%)
- EPS: X.XX元
- 经营现金流: XXX亿

Update sequence:

  1. Drop Q[X] actuals into historical columns
  2. Verify sum checks (quarterly sum = annual)
  3. Update LTM (Last Twelve Months) calculations
  4. Check annual-to-quarter relationships

Step 3: Roll Forward Estimates

Revenue projections:

  • Update growth rates based on Q[X] performance
  • Consider:
  • Order backlog changes
  • New product ramp
  • Market share gains/losses
  • Capacity expansion

Margin projections:

  • Update gross margin based on actual trend
  • Update opex ratios based on actual leverage
  • Flag structural changes (input costs, pricing power)

Balance sheet projections:

  • Update working capital assumptions
  • Update debt schedule if refinancing occurred
  • Update CapEx plans

Step 4: Update Valuation

Market data refresh:

python
get_quote(ticker)  # Current price, PE, PB

Valuation metrics to update:

  • Current stock price and change
  • Trading multiple (PE, PB, EV/Sales, EV/EBITDA)
  • 52-week range position
  • Relative to sector median
  • Implied growth vs historical

DCF updates (if applicable):

  • Roll forward projections by one year
  • Update market data inputs (beta, risk-free rate, shares)
  • Update terminal growth if outlook changed

Step 5: Flag Material Changes

Change significance assessment:

ChangeThresholdAction
Revenue estimate±5%Update model, note driver
Net income estimate±10%Update model, revise target price
Margin estimate±100 bpsUpdate model, assess sustainability
CapEx estimate±20%Update model, check FCF impact
Multiple assumption±1xSensitivity check, document rationale

Red flags to highlight:

  • Revenue growth deceleration >500 bps
  • Margin compression >300 bps
  • Working capital deterioration
  • Debt increase >20% vs prior forecast
  • Management guidance lower than consensus

Step 6: Document Changes

Model update memo:

[公司名称]([代码])模型更新 [Date]

一、更新内容
   [Bullet list of changes made]

二、关键变动
   Revenue FY25E: [Old] → [New] ([Change]%)
   Net Income FY25E: [Old] → [New] ([Change]%)
   Driver: [Explanation]

三、估值影响
   新目标价: ¥XX.XX (之前 ¥XX.XX)
   调整幅度: +X% / -X%
   调整逻辑: [Brief rationale]

四、后续关注
   - [Next catalyst]
   - [Key metric to monitor]
   - [Risk factor]

Step 7: QC Checklist

Before finalizing:

  • [ ] All Q[X] actuals sourced from 巨潮 PDF or AkShare
  • [ ] Historical data matches reported figures exactly
  • [ ] All formulas intact (no broken references)
  • [ ] LTM calculations updated
  • [ ] Forward estimates reflect new information
  • [ ] Valuation inputs refreshed (price, shares, multiples)
  • [ ] Target price recalculated
  • [ ] Change log complete
  • [ ] Cell comments added for new inputs
  • [ ] Model balances correctly

China-Specific Update Considerations

Earnings Season Timing

ReportDeadlineTypical Release Window
Q1季报Apr 30Apr 1-30
中报Aug 31Aug 1-31
Q3季报Oct 31Oct 1-31
年报Apr 30Jan-Apr

Model update priority:

  • Annual report: Complete overhaul of historicals
  • Semi-annual: Major update, adjust full-year estimates
  • Quarterly: Incremental update, verify full-year trajectory

业绩预告 Integration

If company issued 业绩预告:

  • Use as directional signal before formal report
  • Adjust estimates if variance >20% from prior
  • Flag for detailed update when formal report arrives

Consensus Management

  • Update consensus assumptions based on new information
  • If company guidance differs from consensus, flag divergence
  • Note if management commentary suggests estimate revision

Regulatory Changes

Monitor for:

  • Tax rate changes (高新技术企业 reclassification)
  • Accounting standard updates
  • Industry-specific regulation impacts on assumptions
  • Dividend policy changes (影响 DCF terminal value)

Quality Checks

Before delivering:

  • [ ] All new actuals traceable to source
  • [ ] Formulas verified (no hardcodes in calculations)
  • [ ] Estimates logically consistent with new data
  • [ ] Valuation update reflects current market conditions
  • [ ] Changes documented in update memo
  • [ ] Model passes basic QC (balance checks, sum checks)
Data Source Mode Switch: Set env var IFIND_DATA_SOURCE_MODE to control data source preference. - ifind-only (strict): Use iFind only, error if unavailable - ifind-fallback (default): iFind preferred, fallback to AkShare - akshare-only: Skip iFind, use AkShare only

タスクに関連する項目

同じリポジトリから

関連する Skills

すべての Skills
jwangkun
コミュニティ

china-catalyst-calendar

Track upcoming catalysts for A-share coverage universe — earnings dates, regulatory announcements, sector conferences, product launches, and policy events relevant to Chinese equities. Adapted from the original catalyst-calendar skill for A-share market conventions. Triggers on "A股催化剂日历", "事件日历", "财报日历", "earnings calendar A-share", "upcoming catalysts China", or "what's coming up for [company]".

導入数
1
GitHub Stars
840
更新日
9月14日
jwangkun
コミュニティ

china-cim-builder

Structure and draft Confidential Information Memorandums (CIMs) for A-share sell-side M&A processes. Organizes company information into professional, investor-ready documents with Chinese market context. Triggers on "A股CIM", "保密信息备忘录", "CIM China", "confidential memorandum A-share", or "offering memorandum [company]".

導入数
1
GitHub Stars
840
更新日
9月14日
jwangkun
コミュニティ

china-competitive-analysis

Competitive landscape analysis for A-share companies and Chinese industries. Maps competitors, compares positioning, and assesses relative strengths. Adapted from the original competitive-analysis skill for Chinese market context. Triggers on "A股竞争格局", "行业竞争分析", "competitive landscape China", "competitive analysis A-share", or "[company] competitors China".

導入数
1
GitHub Stars
840
更新日
9月14日
jwangkun
コミュニティ

china-datapack-builder

Build professional data packs for A-share investment banking deals — M&A due diligence, IC materials, company profiles. Adapted from the original datapack-builder skill for Chinese data sources (AkShare, 巨潮, exchange filings) and CAS accounting. Triggers on "A股数据包", "投行数据包", "datapack China", "due diligence pack A-share", or "IC data pack [company]".

導入数
1
GitHub Stars
840
更新日
9月14日