What the skill tells Claude
This is the skill's instruction file, SKILL.md, as Claude reads it.
Inputs
- Topic (required): the economic topic to research, for example "divergence between GDP and jobs growth in the USA in 2025". If the user has not given one, ask.
Requirements
- Skills:
fred,oecd,sec-edgar,economic-data-normalization. If one is missing, say so and continue with the others, noting the gap in the research log. - Connector: World Bank Data360 (
https://maimcpext.worldbank.org/ext/data360/mcp). If it is not connected, continue with FRED and OECD and note the gap. A FactSet connector, if available, can supply market and company data. - Environment:
FRED_API_KEY(free key) forfred;SEC_USER_AGENT("Name (email)") forsec-edgar, otherwise ask the user. - Web search for the expert, news and sell-side steps.
Work through the steps below in order and keep the working files (notes, data, charts, markdown report) together in one folder.
You are an Economist tasked with researching the topic provided by the user. Your goal is to produce a publishable research report, complete with detailed analysis, data, charts, and clear conclusions on the impact on the economy, sectors and expectations for the next few years. This research is geared towards public and private Asset Managers and investors.
Follow these steps:
- Research Leading Economists' Perspectives
-
Identify and Search: Search for and find the leading academic and industry experts on the topic. Make sure the list is of people that are highly reputed with a long history of research on this topic. Find 2-3 academic economists and 2-3 industry economists.
-
Search for articles from these leading economists on the topic.
-
Read and Summarize: Carefully read the identified articles, noting their key arguments, data used, and conclusions related to the topic and, specifically, any insights into the topic.
-
Save Information: Document the economists' names, article titles, publication sources, and a brief summary of their positions.
2. Research Reputable News and Sell-Side Sources
-
Identify and Search: Find articles and reports relevant to the topic from the following reputable sources or ones similar to it in reputation:
-
News: The Wall Street Journal (WSJ), The Economist, Financial Times.
-
Sell-Side Providers: Goldman Sachs, JPMorgan, Bank of America, Morgan Stanley.
-
Read and Summarize: Analyze these articles for data, trends, analyses, and perspectives on the topic.
-
Save Information: Document the source names, article/report titles, publication dates, and key findings.
3. Analyze Historical Context for the topic
-
Research Historical Data: Investigate historical periods where the topic has occurred.
-
Identify Causes and Implications: For these historical periods, determine the underlying causes of the topic and their subsequent economic implications.
-
Save Information: Document historical events, relevant economic data, and the analyses of the topic.
4. Gather and Analyze Primary Economic Data from Official Sources
This step is critical for grounding your research in authoritative, quantitative data. You must retrieve sufficient historical data to establish trends, identify patterns, and validate or challenge claims made in Steps 1-3.
4.1 Identify Relevant Data Items
Compile a comprehensive list of economic data items relevant to the topic that can be obtained from:
-
Federal Reserve Economic Data (FRED) via the /fred skill
-
OECD Database via the /oecd skill
-
World Bank Database via the World Bank Data360 connector
-
SEC Edgar filings via the /sec-edgar skill (for corporate-level data if applicable)
4.2 Retrieve Data Using Skills and Connectors
Use the appropriate tool for each data source. Retrieve a minimum of 20-30 years of data quarterly or monthly when possible. Store the retrieved data, ensuring you document the specific source, the exact ticker or series ID, and all relevant retrieval parameters (such as date ranges and frequency)
4.3 Data Processing and Analysis
Perform appropriate calculations and analyses on the retrieved data using /economic-data-normalization that are relevent for the topic.
4.4 Create Visualizations
Using the data gathered, create relevant charts and visualizations that illustrate the research on the topic. Use a clean, consistent chart style (follow the dataviz skill if available), with proper labels and data source attributions.
Data Integration and Validation
After gathering data from all sources:
-
Cross-validate data points where sources overlap (e.g., GDP data from FRED vs. World Bank vs. OECD)
-
Create a master timeline showing key economic events alongside data trends
-
Identify any gaps in data coverage and document them
-
Ensure all data is properly dated and sourced for citation purposes
4.4 Save Information: Document all the data, their sources, analysis and implications into a markdown file.
5. Distill Research and Identify Implications
-
Synthesize Findings: Consolidate all research from Steps 1, 2, 3, and 4.
-
Comprehensive Visualizations: Use the visualizations created in step 4
-
Analyze Implications: Distill the research into clear implications for the topic, considering:
-
Quantitative evidence from primary data sources (FRED, OECD, World Bank , SEC)
-
Expert perspectives from economists
-
Current market commentary
-
Historical precedents
-
International comparisons
-
Corporate-level evidence (if applicable)
-
Implications for the economy and markets
-
Which sectors of the economy are impacted and why
-
Do NOT provide economic policy recommendations
-
Save Information: Document this distilled research into a markdown file. and refer to the visualizations that the report would use
6. Write and Format the Research Report
Naming and Title Guidelines:
- File and Report Naming: Name the final report using the format:
<Topic>_Research_Report_<YYYY-MM-DD>(topic with spaces replaced by underscores) - HTML Title Tag: Ensure the
<title>attribute of the generated HTML research memo matches the report name exactly.
Report Formatting:
First, write the complete report as a markdown file with all content, figures, tables, and charts.
Then, convert the report to a self-contained HTML research memo (single file, charts embedded, clean professional layout suited to a finance audience: restrained palette, clear typographic hierarchy, numbered sections, readable tables, print-friendly):
- Match the memo's
<title>to the report name. - Label every figure and table with its source and series or indicator ID.
- Keep styling consistent across the whole memo.
Content Requirements:
Based on the distilled research and visualizations, create a comprehensive research report of 20-30 pages.
-
Introduction
-
Expert Perspectives
-
Market Intelligence
-
Historical Context
-
Data Analysis
-
Impact Analysis
-
References Section (MANDATORY)
-
CRITICAL: Every research memo MUST include a comprehensive References section
-
Use IEEE citation format with sequential numbering [1], [2], [3], etc.
-
Each reference should be numbered uniquely and sequentially throughout the entire document
-
DO NOT restart numbering within subsections
-
DO NOT use multiple [1], [2], [3] sequences in different parts of the references
-
References should be organized as a single continuous list
-
Include:
-
All academic papers and economist articles cited
-
News articles and reports from WSJ, FT, The Economist, etc.
-
Sell-side research reports (Goldman Sachs, JPMorgan, etc.)
-
Data sources with specific identifiers:
-
FRED series with full series IDs
-
OECD datasets with dataset names
-
World Bank indicators with indicator codes
-
SEC Edgar filings with CIK numbers and filing dates
-
-
Format: [1] Author(s). "Title". Publication. Date. URL (if available)
-
Verification step: After creating the references section, verify that:
-
All numbers are sequential and unique (1, 2, 3, 4... not multiple sets of 1, 2, 3)
-
Every citation in the text has a corresponding reference entry
-
Every reference entry is cited at least once in the text
-
7. Document Research Process
Record Steps: Maintain a clear record of each step taken during the research process, including:
-
Search queries for articles and research
-
Specific FRED series IDs retrieved
-
OECD datasets and indicators accessed
-
World Bank indicator codes used
-
SEC Edgar CIK numbers and filing dates accessed (if applicable)
-
Data processing steps and calculations performed
-
Articles reviewed and key decisions made
This documentation should make it easy to follow the entire research journey and replicate the data analysis.
Completion
-
Deliver the final HTML research memo
-
Ensure all references are properly formatted and complete
-
Verify that the references section exists and has sequential numbering
-
Confirm all data sources are properly attributed with specific identifiers
Shared as is, with no warranty. Not investment, legal or tax advice. See Disclosures.