financial-report-analysis

In-Depth Financial Statement Analysis (Financial Report Analysis)

Target Role

Role: Corporate Client Manager, Credit Approval Officer, Risk Manager
Use Case: Pre-loan due diligence financial analysis, annual post-loan review, large credit approval, restructuring/extension assessment
Output Purpose: Generate a structured financial health assessment report to provide data support for credit decisions and investment decisions
Decision Level: Core reference material for credit approval, risk level high, requires review by a credit approval officer
Execution Frequency: Executed once before each credit application, and at least once per year after the loan

Data Sources

Required Data

Data Item	Source	Acquisition Method	Sensitivity Level
The three financial statements	User upload/public disclosure	File reading (PDF/Excel)	Internal
Notes to the financial statements	User upload/public disclosure	File reading	Internal
Audit report	User upload/public disclosure	File reading	Internal
Industry financial benchmarks	Wind/Tonghuashun	API/data subscription	Public
Historical financial analysis reports	Bank's imaging archive system	API: /api/reports/list	Internal

Data Masking Rules

Customer ID number: show the first 3 and last 4 digits, with the middle replaced by *
Bank account number: show only the last 4 digits
Customer confidential business information (such as key customer lists and supplier details): use "a certain customer/a certain supplier" instead in the report
Sensitive financial data (such as undisclosed performance): use only in internal reports and do not disclose externally
Related-party information: mark full names in internal reports, and use "Related Party A/B/C" in external reports

Degradation Strategy

If the audit report is unavailable: mark "audit opinion not verified", continue the analysis based on the financial statements, but lower the credit rating
If the notes to the financial statements are missing: mark "notes not provided, some account details cannot be verified", and analyze only the data in the three statements
If industry benchmark data is unavailable: mark "industry benchmarking data not obtained", and perform only vertical trend analysis
If historical financial reports are unavailable: mark "no historical reports for comparison", and perform only static analysis for the current period
If financial data is available for only 1 year: mark "insufficient data, trend analysis unavailable", and perform only single-period analysis
If the system is unavailable (Wind/Tonghuashun): use the most recently downloaded industry benchmark data (mark the data date)

Workflow

Step 0: Data Confirmation and Verification

List input parameters: enterprise name, stock code (if any), reporting period, analysis scenario (pre-loan due diligence/post-loan review/risk warning). Confirm the data time range (last 3-5 years) and accounting standards (Chinese Accounting Standards for Business Enterprises/International Financial Reporting Standards). Run scripts/validate_financial_report.py Check the completeness of input parameters and the cross-check relationships among the three financial statements.

✅ Verification passed → proceed to Step 1
❌ Verification failed → return the list of missing items and ask the user to supplement (e.g., "Please provide complete three financial statements for 2021-2023")

📋 Data source:user_upload(financial statements uploaded by the user)

Step 1: Data Collection and Quality Assessment

Confirm the source of financial data and perform preliminary data quality screening:

Type of audit opinion (standard unqualified/with emphasis of matter paragraph/qualified/adverse/disclaimer of opinion)
Whether accounting policies have changed (quantify the impact of changes)
Whether the scope of consolidated financial statements has changed
Identification of significant changes in accounting estimates

If the data is incomplete, proactively request from the user:

Complete three financial statements for the last 3-5 years (balance sheet, income statement, cash flow statement)
Notes to the financial statements (details of significant accounts)
Audit report (especially focus on the audit opinion)

The data quality assessment step must not be skipped. If the audit opinion triggered is a non-standard opinion, it must be prominently noted at the beginning of the report.

✅ Data complete → proceed to Step 2
⚠️ Partially missing → mark the missing items, continue the analysis but lower the confidence level
❌ Severely missing data (only 1 year or missing core financial statements) → output a "Insufficient data, unable to complete the analysis" report

📋 Data source:user_upload or reference(public disclosure)

Step 2: In-Depth Profitability Analysis

Calculation of core profitability indicators(Refer to references/financial-indicators-calculation.md):

Gross profit margin, core profit margin, net profit margin
ROE、ROA、ROIC

Profit quality assessment:

Core profit proportion (>80% is excellent)
Proportion of non-recurring gains and losses
Cash collection ratio (>1 is excellent)
Net cash ratio (>1 is excellent)

Profitability decomposition:

Attribution of gross profit margin changes: volume-price decomposition
Attribution of expense ratio changes
Judgment on the sustainability of profit margin changes

No profit quality assessment step may be skipped. All numbers must show the calculation process. If the cash collection ratio or net cash ratio < 0.5 for 2 consecutive years, it must be marked as "low profit quality".

✅ Data complete → proceed to Step 3
⚠️ Some data missing → mark "a certain indicator cannot be calculated", continue the analysis

📋 Data source:context(financial statements uploaded by the user)

Step 3: In-Depth Balance Sheet Analysis

Asset quality assessment:

Accounts receivable: aging analysis + adequacy of bad debt provisions + concentration of top five customers
Inventory: inventory age analysis + write-down provisions + matching with revenue
Goodwill: proportion of net assets + reasonableness of impairment testing assumptions
Fixed assets: depreciation policy + novelty rate + capacity utilization
Construction in progress: reasonableness of construction period + whether there are long-term outstanding items

Liability structure analysis:

Interest-bearing debt ratio, short-term and long-term debt ratio
Operating liabilities vs financial liabilities
Off-balance-sheet liabilities (guarantees, leases, notes payable, etc.)

Asset-liability matching:

Current ratio & quick ratio
Working capital trend
Maturity mismatch risk

Do not skip any core asset analysis step. If other receivables/net assets > 20%, the risk of "possible related-party fund occupation" must be flagged.

✅ Complete data → Proceed to Step 4
⚠️ Missing notes → Flag "notes not provided; details of certain accounts cannot be verified"

📋 Data source:context(financial statements uploaded by the user)

Step 4: In-depth cash flow analysis

Operating cash flow analysis:

Analysis of the difference between operating cash flow and net profit (item-by-item reconciliation)
Free cash flow = operating cash flow - capital expenditure

Investing cash flow analysis:

Scale and direction of capital expenditure
Capital expenditure vs depreciation and amortization
Reasonableness of M&A expenditure

Financing cash flow analysis:

Choice of financing method (equity vs debt)
Dividend payout ratio and stability
Share repurchase behavior

Cash flow combination signal identification(See references/cashflow-pattern-guide.md)

Do not skip cash flow quality analysis. If operating cash flow has been negative for 3 consecutive years and interest-bearing debt continues to increase, red line R2 must be triggered.

✅ Complete data → Proceed to Step 5
⚠️ Cash flow statement missing → Flag "cash flow statement not provided; cash flow quality cannot be analyzed"

📋 Data source:context(financial statements uploaded by the user)

Step 5: DuPont analysis and driver decomposition

Classic three-factor DuPont analysis:

ROE = 净利率 × 总资产周转率 × 权益乘数

Five-factor DuPont decomposition(more granular):

ROE = 税负系数 × 利息负担系数 × 经营利润率 × 资产周转率 × 权益乘数

Change attribution analysis:

Year-over-year change in ROE
Contribution of each of the three factors (chain substitution method or difference analysis method)
Identification of core drivers and judgment of sustainability

Do not skip the DuPont analysis step. The calculation process of the chain substitution method must be shown, and the core drivers must be identified.

✅ Complete data → Proceed to Step 6
⚠️ Some data missing → Flag "a certain factor cannot be decomposed"; continue the analysis

📋 Data source:context(financial statements uploaded by the user)

Step 6: Financial early warning and window-dressing identification

Identification of common financial window-dressing methods(See references/fraud-detection-guide.md):

Revenue side: premature revenue recognition, channel stuffing, inflation through related-party transactions
Cost side: expense capitalization, understated depreciation/impairment
Profit side: smoothing of non-recurring gains and losses, fair value manipulation
Cash flow side: packaging financing cash flow as operating cash flow

Key early warning signal scan(3 or more require high vigilance):

Accounts receivable growth rate >> revenue growth rate
Net cash ratio < 0.5 for more than 2 consecutive years
Inventory turnover days continuously lengthening
Goodwill / net assets > 30%
Construction in progress not transferred to fixed assets for a long time
Abnormal increase in other receivables
Frequent changes in accounting policies/estimates
High deposits and high loans

Red line condition self-check(8 one-vote-veto conditions R1-R8): Check the 8 red lines one by one. If any red line is triggered, it must be prominently flagged at the beginning of the report.

Do not skip the red line self-check step. R1-R8 must be checked one by one, with none omitted.

✅ No red line triggered → Proceed to Step 7
⚠️ Trigger 1-2 red lines → Mark red warning, continue analysis
❌ Trigger more than 3 red lines → Mark "Recommend suspending credit decision", proceed to step 7

📋 Data source:context(User-uploaded financial statements)

Step 7: Valuation reference (if needed)

Relative valuation:

PE (TTM & Forward) vs historical percentile vs industry average
PB vs ROE matching
EV/EBITDA cross-company comparability

Absolute valuation (summary framework):

DCF core assumptions: perpetual growth rate, WACC
Free cash flow forecast logic

📋 Data source:reference(Industry valuation database)

Step 8: Verification of the articulation relationships among the three financial statements

Perform verification of the articulation relationships among the three financial statements:

Balance sheet ending balance minus beginning balance = cash flow statement ending cash
Income statement net profit matches changes in balance sheet equity
Cash flow statement operating cash flow is consistent with income statement net profit adjustments

If the articulation relationships are inconsistent, it must be marked "The data in the three financial statements is inconsistent, please verify the data source".

✅ Articulation relationships consistent → Proceed to step 9
❌ Articulation relationships inconsistent → Mark inconsistent items and require the user to verify

📋 Data source:context(User-uploaded financial statements)

Step 9: Generate financial analysis report

Use the assets/financial-report-template.md template to generate a structured financial health assessment report.

The report must contain the following sections:

Company overview and analysis framework
Profitability analysis (including drivers breakdown, quality assessment, sustainability judgment)
Asset quality analysis (core assets, accounts receivable, inventory, goodwill)
Liabilities and solvency (structure analysis, debt repayment pressure, maturity matching)
Cash flow analysis (overview of the three major cash flows, operating cash flow quality, free cash flow)
DuPont analysis (ROE three-factor breakdown, change attribution, peer benchmarking)
Financial warning scan (list of abnormal signals, articulation relationship verification, signs of earnings management)
Comprehensive evaluation and recommendations (financial health, core strengths and risks, tracking indicators, decision reference)

All data should be labeled with: data source + reporting period + whether audited. The report must include a disclaimer at the end (citing shared/disclaimer-template.md)。

📋 Confirmation mechanism:inform(Notify the account manager for confirmation after generation)

Output Format

Use the assets/financial-report-template.md template. The report must contain the following sections:

Company overview and analysis framework
Profitability analysis (including drivers breakdown, quality assessment, sustainability judgment)
Asset quality analysis (core assets, accounts receivable, inventory, goodwill)
Liabilities and solvency (structure analysis, debt repayment pressure, maturity matching)
Cash flow analysis (overview of the three major cash flows, operating cash flow quality, free cash flow)
DuPont analysis (ROE three-factor breakdown, change attribution, peer benchmarking)
Financial warning scan (list of abnormal signals, articulation relationship verification, signs of earnings management)
Comprehensive evaluation and recommendations (financial health, core strengths and risks, tracking indicators, decision reference)

All data should be labeled with: data source + reporting period + whether audited.

Downstream compatibility: This output can be parsed and used by submit-credit-application . Key fields include:

Financial health rating (string: Excellent/Good/Fair/Warning/Danger)
Core financial indicators (table: ROE, gross margin, net margin, debt-to-asset ratio, DSCR, etc.)
Red line trigger list (array: if triggered, list R1-R8 numbers)
Audit opinion type (string: Unqualified/With Emphasis of Matter/Qualified/Adverse/Disclaimer)

Disclaimer: The end of the report must cite shared/disclaimer-template.md, ensuring it includes necessary statements such as "This analysis does not constitute investment advice".

Constraints

Data traceability: All conclusions must clearly indicate the data source (annual report/quarterly report/audit report/notes), citing specific page numbers or accounts
Explicit labeling of uncertainty: For items with missing data, inconsistent standards, or inability to verify, they must be explicitly labeled as "to be verified/data missing"; speculative filling is prohibited
Timeliness requirements: All financial data must indicate the reporting period; analysis conclusions must cover the past 3 years (credit scenarios require the past 5 years); cross-period comparisons must use consistent standards
Quantitative conclusions preferred: All qualitative risk judgments (e.g., "profitability declining") must be accompanied by specific quantitative indicators (e.g., "gross margin cumulative decline of 8.2pct over 3 years"); conclusions based solely on qualitative descriptions are prohibited
Neutral and prudent stance: Remain prudent toward optimistic data, provide counter-verification for positive interpretations, and avoid management narratives dominating analysis conclusions
Three-statement articulation verification: Balance sheet ending minus beginning difference = cash flow statement ending cash; income statement net profit matches balance sheet equity changes
Industry benchmarking: Core indicators such as ROE, gross margin, and debt-to-asset ratio must provide industry average references
Skipping steps prohibited: No steps in the Workflow may be skipped, especially data quality assessment, earnings quality assessment, red line self-check, and articulation verification
Mandatory red line enforcement: If any red line (R1-R8) is triggered, it must be prominently marked at the beginning of the report, and suspension of credit decision-making must be recommended

Financial compliance red lines (one-vote veto)

The following situations must be marked in the report summary withred warning, and suspension of credit decision-making must be recommended, pending re-evaluation after verification:

R1: Audit opinion is adverse opinion or disclaimer of opinion — Reliability of financial data is questionable; it must not be used as a basis for credit
R2: Operating cash flow has been negative for 3 consecutive years, and interest-bearing debt continues to increase — The enterprise lacks self-sustaining cash generation ability and relies on borrowing new to repay old
R3: Other receivables / net assets > 30%, and the main counterparties are related enterprises — There may be related-party fund occupation
R4: Monetary funds balance > 20% of total liabilities but at the same time a large amount of short-term interest-bearing debt is borrowed (high deposits and high loans) — Funds may be restricted or misappropriated
R5: Accounts receivable growth rate exceeds operating revenue growth rate by more than 50 percentage points for 2 consecutive years — There may be inflated revenue or deteriorating collections
R6: Major construction in progress has not been transferred to fixed assets for a long time (>3 years), and there is no reasonable explanation of the construction cycle — Expenses may be hidden or the project may be stalled
R7: Goodwill as a proportion of net assets > 50%, and the key assumptions of the impairment test are clearly optimistic — Significant impairment risk exists
R8: Within the past 1 year, the actual controller or major shareholders pledged a high proportion of equity (> 80%) and no remedial measures were taken — The actual controller's capital chain is tight

Analysis Principles

Substance over form: Look through accounting treatment to the operating substance; focus on cash flow rather than only profit
Horizontal and vertical comparison: Look at trend changes vertically and benchmark against peers horizontally; identify problems through comparison
Anomaly-driven: Focus on accounts with abnormal fluctuations and analyze the business logic behind them
Conservative and prudent: Take a prudent attitude toward optimistic data and remain highly sensitive to risk signals
Cross-footing verification: Cross-verify the reasonableness of data through the cross-footing relationships among statements
Quantified conclusions: All qualitative judgments must be supported by quantitative data

Credit Scenario-Specific Focus Areas

This skill serves both investment analysis and credit underwriting scenarios. When used for bank pre-loan/post-loan due diligence, the following dimensions have higher priority:

Dimension	Credit Focus	Typical Risk Signals
Repayment capacity	Whether operating cash flow can cover annual principal and interest payments (DSCR ≥ 1.2)	DSCR < 1.0 for 2 consecutive years; free cash flow is negative
Asset realisability	Liquidity of the book assets corresponding to collateral and realisation discounts	Accounts receivable aging > 1 year accounts for more than 30%; inventory write-down provision is insufficient
Related-party fund occupation	Whether there is any transfer of funds to related parties in other receivables or prepayments	Other receivables/net assets > 20%; the top five prepayment counterparties are related parties
Financial leverage	Whether interest-bearing debt/EBITDA exceeds the industry tolerance上限	Interest-bearing debt/EBITDA > 5× or interest-bearing debt ratio > 60%
Earnings manipulation	Profit quality (net cash ratio) and revenue authenticity (cash collection ratio)	Net cash ratio < 0.5 for 2 consecutive years; receivables growth exceeds revenue growth by more than 50pct
Stress resilience	Whether debt-servicing capacity can still be supported under extreme scenarios (revenue down 30%)	Under stress testing, interest coverage ratio < 1.5

Audit Trail

After each financial statement analysis is completed, generate an audit log audit/{企业简称}_{日期}_audit.json:

{
  "skill_name": "financial-report-analysis",
  "skill_version": "2.0.0",
  "execution_time": "2026-05-05T10:30:00+08:00",
  "input_params": {
    "company_name": "XX股份",
    "stock_code": "600XXX",
    "report_period": "2023年报",
    "analysis_depth": "全面(Step 0-9)",
    "analysis_scenario": "贷前尽调"
  },
  "operator": "客户经理姓名(工号:XXX)",
  "steps": [
    {
      "step": "数据确认与验证",
      "executor": "ai",
      "data_source": {"type": "user_upload", "files": ["2023年报.pdf"]},
      "result": "pass",
      "validation_script": "validate_financial_report.py"
    },
    {
      "step": "数据收集与质量评估",
      "executor": "ai",
      "data_source": {"type": "user_upload", "files": ["2023年报.pdf"]},
      "result": "pass",
      "audit_opinion": "标准无保留意见"
    },
    {
      "step": "盈利能力深度分析",
      "executor": "ai",
      "data_source": {"type": "context"},
      "result": "pass"
    },
    {
      "step": "财务预警扫描",
      "executor": "ai",
      "data_source": {"type": "context"},
      "result": "pass",
      "red_flags_triggered": []
    },
    {
      "step": "三表勾稽关系验证",
      "executor": "ai",
      "data_source": {"type": "context"},
      "result": "pass",
      "reconciliation_status": "一致"
    },
    {
      "step": "生成财务分析报告",
      "executor": "ai",
      "confirmation": {"type": "inform", "notified_to": "客户经理姓名", "notified_at": "2026-05-05T10:35:00+08:00"},
      "result": "pass"
    }
  ],
  "warnings": ["部分附注数据缺失"],
  "references_used": ["references/financial-indicators-calculation.md", "references/fraud-detection-guide.md"]
}

Audit log retention period: At least 3 years.

Gotchas

#1: Looking only at profit and not at cash flow

Symptoms: The income statement shows profit growth, but operating cash flow is negative
Cause: Failure to strictly implement earnings quality assessment (cash collection ratio/net cash ratio)
Solution: Each analysis must calculate the cash collection ratio and net cash ratio; if < 0.5 for 2 consecutive years, mark "low profit quality"

#2: Lack of industry benchmarking

Symptoms: The analysis report only makes vertical trend comparisons and does not benchmark against peers
Cause: Industry benchmark data was not obtained or this step was ignored
Solution: Industry averages must be obtained; core indicators such as ROE, gross margin, and debt-to-asset ratio must be marked "Industry average: X%"

#3: Red-line signals are drowned out in the main text

Symptoms: 8 red-line conditions triggered, but not prominently flagged in the report summary
Cause: The red-line self-check was not performed after Step 6 was completed
Solution: After each analysis is completed, the 8 red lines must be self-checked, and any trigger must be marked with a red warning in a prominent position in the summary

#4: The cross-check relationships among the three statements were not verified

Symptoms: The balance sheet and cash flow statement data are inconsistent, but this was not detected
Cause: The three-statement cross-check verification was not performed
Solution: Balance sheet ending balance - beginning balance = cash flow statement ending cash; income statement net profit matches changes in balance sheet equity

#5: Note disclosures were ignored

Symptoms: Only the three main statements were analyzed, without attention to major matters, contingent liabilities, or commitments in the notes
Cause: The user was not asked to provide the notes, or the notes were not read carefully
Solution: The notes must be requested, with key attention to related-party transactions, contingent liabilities, and changes in accounting policies

Examples

Example 1: Pre-loan due diligence financial analysis

User input:

请分析XX制造2023年年报,用于贷前尽调。授信额度5000万,期限1年。

Skill execution process:

Data confirmation → Verify completeness of input parameters, run validate_financial_report.py
Data collection → Obtain the three statements for 2021-2023 + audit report + notes
Data quality assessment → Audit opinion: standard unqualified, no major changes in accounting policies
Profitability analysis → Gross margin 28.5% (industry average 25.2%), net cash ratio 1.2 (good earnings quality)
Balance sheet analysis → Interest-bearing debt ratio 35% (< 40% safety threshold), current ratio 1.8
Cash flow analysis → Operating cash flow remained positive, free cash flow turned negative in 2023 (capacity expansion investment)
DuPont analysis → ROE 15.2%, core driver: net profit margin improvement (not financial leverage)
Red-line scan → Not triggered
Cross-check verification → The cross-check relationships among the three statements are consistent
Generate report → Financial health rating: Good

Output points:

Core conclusion: Good earnings quality, healthy cash flow, with repayment capacity for a CNY 50 million credit facility
Risk warning: Free cash flow was negative in 2023 (capacity expansion investment); attention should be paid to production launch progress
Recommendation: The credit facility may be approved, with a recommendation to add a factory mortgage

Example 2: Risk warning financial analysis

User input:

XX贸易公司近期有逾期记录,请分析其2023年财报,评估风险。

Skill execution process:

Data confirmation → Verify completeness of input parameters
Data collection → Obtain the three statements for 2021-2023 + audit report
Data quality assessment → Audit opinion: unqualified opinion with emphasis of matter paragraph (material uncertainty related to going concern)
Profitability analysis → Gross margin 5.2% (industry average 8.5%), net cash ratio 0.3 (low quality of earnings)
Balance sheet analysis → Interest-bearing debt ratio 68% (> 60% red line), current ratio 0.9 (< 1.5)
Cash flow analysis → Operating cash flow has been negative for 2 consecutive years, financing cash flow was positive (borrowing new to repay old)
DuPont analysis → ROE -2.5% (loss), core drag: operating profit margin turned negative
Red-line scan → Triggered R2 (operating cash flow negative for 3 consecutive years + increase in interest-bearing debt) and R5 (receivables growth exceeded revenue growth by 60 percentage points)
Cross-check verification → The cross-check relationships among the three statements are consistent
Generate Report → Financial Health Rating: Warning

Output Key Points:

⚠️ Red Warning: Two red lines, R2 and R5, triggered; it is recommended to suspend the credit decision
Core Conclusion: The enterprise lacks cash-generating capability, relies on borrowing new to repay old, and receivables collection is deteriorating
Recommendation: Reassess after verifying the aging of receivables and related-party fund occupation

Out of Scope

This Skill does not handle credit approval decisions and only provides financial analysis data support
This Skill does not generate credit application reports (please use the submit-credit-application Skill)
This Skill does not directly modify customer data or submit credit applications
This Skill does not handle financial analysis for personal credit/retail business
This Skill does not conduct market research or industry macro analysis (please use the credit-industry-analysis Skill)
This Skill does not perform automated financial data scraping (users must upload or manually provide it)
If the user requests any of the above, clearly inform them and recommend the appropriate Skill or contact the corresponding department

About this skill

Target Role

Data Sources

Required Data

Data Masking Rules

Degradation Strategy

Workflow

Step 0: Data Confirmation and Verification

Step 1: Data Collection and Quality Assessment

Step 2: In-Depth Profitability Analysis

Step 3: In-Depth Balance Sheet Analysis

Step 4: In-depth cash flow analysis

Step 5: DuPont analysis and driver decomposition

Step 6: Financial early warning and window-dressing identification

Step 7: Valuation reference (if needed)

Step 8: Verification of the articulation relationships among the three financial statements

Step 9: Generate financial analysis report

Output Format

Constraints

Financial compliance red lines (one-vote veto)

Analysis Principles

Credit Scenario-Specific Focus Areas

Audit Trail

Gotchas

#1: Looking only at profit and not at cash flow

#2: Lack of industry benchmarking

#3: Red-line signals are drowned out in the main text

#4: The cross-check relationships among the three statements were not verified

#5: Note disclosures were ignored

Examples

Example 1: Pre-loan due diligence financial analysis

Example 2: Risk warning financial analysis

Out of Scope