Financial Report Analyzer
Automatically extract key data from listed-company financial reports, compute ratios, generate charts, and produce professional analysis reports.
The skill operates within reasonable scope: it only parses financial report files, calculates ratios, and generates charts, using the dedicated execute_skill_script_file tool, with no external network requests or sensitive data exfiltration. However, no user confirmation mechanism exists; scripts are executed automatically by the main agent without user confirmation for key steps. Least privilege was not explicitly addressed. No rollback or recovery mechanism. Score 10: deductions for missing user confirmation, incomplete permission scoping, and lack of rollback.
The workflow is clear, scripts have well-defined responsibilities, and the key path is reproducible. Error handling is limited: extract_financials.py returns errors for missing files, read failures, and empty content, but calculate_ratios.py and generate_charts.py degrade gracefully to default values for invalid input, with insufficient error feedback. Dependencies like pdfplumber and matplotlib are mentioned but no installation instructions. Static assessment cannot verify execution, so score 7: deductions for lack of tests, weak failure feedback, and incomplete dependency guidance.
The skill clearly targets Chinese listed company financial report analysis with a defined audience and scenarios. Capability boundaries are partially defined for PDF and text files but do not state limitations like scanned PDFs or non-standard reports. Trigger conditions are clear from the description and workflow. Environment fit: Chinese support included, no dependency on overseas services, accessible from mainland China. Score 9: deductions for unclear non-fit boundaries and trigger precision.
Information architecture is well-layered: SKILL.md for workflow, references for metrics and framework, and templates for output. Dependency and installation notes are missing (pdfplumber, matplotlib). No versioning or changelog, maintenance responsibility unclear. License is MIT, provenance clear. Worked example present, no FAQ. Known limitations mentioned but incomplete. Score 9: deductions for missing install notes, versioning, and maintenance ownership.
The skill has clear goals and benefits: automated data extraction, ratio calculation, chart generation, and report production. Output completeness depends on the LLM's depth of analysis, and the HTML report requires the platform for viewing; direct usability unverified. Marginal value over manual analysis is clear, but cost-benefit not quantified. Static review cannot validate output quality, so score 6: deductions for no executed verification of output and limited comparative benefit evidence.
Source scripts are auditable, but no test coverage for the skill's key paths. Calculation logic can be statically audited, but extraction accuracy depends on regex matching of financial reports, which is unverified. No third-party execution evidence or CI for this skill. Score 4: deductions for no test suite, no independent execution validation, and unreproducible key claims.
- Skills scripts execute automatically and can read arbitrary file paths; ensure the file path input is trusted and enforce least-privilege settings.
- Chart generation relies on matplotlib and CJK fonts; documentation lacks installation guidance, which may cause garbled Chinese characters.
- Financial data extraction relies on regex and may be inaccurate due to report formatting variations; recommend manual verification.
What does this skill do, and when should you use it?
This skill is designed for in-depth analysis of financial reports from public companies, such as annual and quarterly reports. It parses PDF or text files, automatically extracts core metrics like revenue, net profit, assets and liabilities, calculates key financial ratios (gross margin, net margin, ROE, debt-to-asset ratio, etc.), and generates three visualization charts. Using a built-in analysis framework, the LLM writes seven narrative sections covering profitability, solvency, operational efficiency, and cash flow, which are then rendered into a complete HTML report via a template. The workflow is highly automated, making it suitable for investment research, financial analysis, and due diligence.
- Executes extract_financials.py via execute_skill_script_file to parse PDF files (using pdfplumber) or text files, extracting revenue, net profit, assets, liabilities, and other core values, returning JSON.
- Runs calculate_ratios.py with the extracted JSON to compute 30 key metrics (gross margin, net margin, ROE, debt-to-asset ratio, etc.) and outputs template placeholder keys.
- Executes generate_charts.py to produce three charts: a bar chart of core metrics (financial_overview.png), a horizontal bar chart for profitability (profitability.png), and a donut chart for asset structure (asset_structure.png).
- References financial_metrics.md and analysis_framework.md, and the LLM writes seven analysis sections (including advantage and risk lists) covering profitability, solvency, efficiency, and cash flow.
- Calls html_interpreter with template_path mode, reading report_template.html and automatically combining ratio data, chart URLs, and analysis text to render the final report.
- Terminates with a brief summary, and the report is displayed as a card in the left panel.
- A financial analyst needs to quickly analyze a company's annual report, extract key metrics, and generate a structured report
- An investment advisor wants to provide clients with chart-rich financial report interpretations
- A research team processing multiple quarterly reports and seeking workflow automation
- An individual investor reading a PDF annual report wants AI assistance in understanding financials and risks
- A bank conducting pre-loan due diligence on a corporate borrower needs structured financial analysis
What are this skill's strengths and limitations?
- Highly automated: from PDF extraction to final report, saving significant manual effort
- In-depth analysis: covers profitability, solvency, efficiency, and cash flow dimensions
- Visualization: auto-generates three professional charts for intuitive understanding
- Customizable: templates and reference docs can be modified to adjust report style and focus
- Platform-specific: scripts rely on DB-GPT-specific tools (execute_skill_script_file, html_interpreter), may require adaptation for other Agent platforms
- PDF parsing accuracy is affected by document layout; extracted values may need manual verification
- No test suite; stability across OS/Python versions is not guaranteed
- Only supports PDF and text; no Excel or other spreadsheet input
How do you install this skill?
This skill is part of the open-source DB-GPT project (GitHub: eosphoros-ai/DB-GPT). You need to clone the repository and install DB-GPT. Steps:
- Clone the repo: git clone https://github.com/eosphoros-ai/DB-GPT.git
- Follow the quick start in README (e.g., one-line installer: curl -fsSL https://raw.githubusercontent.com/eosphoros-ai/DB-GPT/main/scripts/install/install.sh | bash)
- Ensure Python 3.10+ and install matplotlib (required by generate_charts.py).
- Start the DB-GPT webserver and the skill will be available.
How do you use this skill?
In the DB-GPT chat interface, ask the AI to analyze a financial report, e.g.:
“Please analyze this annual report: /path/to/your/report.pdf”
The AI will invoke the skill, extract data, compute ratios, generate charts, and produce a report. Specify the file path in your prompt.