FRED economic data
US economic series from the Federal Reserve: GDP, jobs, rates and inflation.
Needs: A free FRED API key, set as FRED_API_KEY.
Projects · AI skills
AI skills for Claude.
11 skills that pull economic data, run attribution and valuation, draft research memos and check marketing documents against SEC and FINRA rules.
I built these to take the repetitive parts out of investment research: fetching the data, cleaning it, attributing performance and checking documents. Each skill is a folder with instructions and scripts that Claude reads when a task calls for it. Download one, or all of them.
Market and economic data
Analysis and methods
Economic data normalization
Outlier handling, seasonal adjustment, smoothing and normalization for economic series.
Needs: Python packages (listed below).
Company valuation
Picks three or four valuation models from a library of 22, runs them and writes the report.
Needs: SEC EDGAR and the web. PitchBook and FactSet where you have access.
Performance attribution
Equity attribution (Brinson-Fachler)
Allocation, selection and interaction effects with Carino multi-period linking, from a holdings file.
Needs: A holdings xlsx or JSON file.
Fixed income attribution (Campisi)
Splits bond returns into income, Treasury, spread and security-specific effects.
Needs: An xlsx with holdings, yield curve and spread curves.
Macro attribution (multi-asset)
Total-portfolio attribution: tactical allocation, manager selection and interaction.
Needs: An xlsx with policy and actual weights and returns.
Research and compliance
Economic deep research
A 20 to 30 page research memo: expert views, 20 to 30 years of official data, charts and IEEE citations.
Needs: The fred, oecd, sec-edgar and economic-data-normalization skills, the World Bank connector and web search.
Marketing compliance review
Reviews marketing documents against SEC, FINRA and GIPS rules and annotates DOCX files with comments.
Needs: python-docx. A first-pass check, not legal advice.
Advisory knowledge
Wealth and estate planning
2026 tax, trust and state-domicile guidance for high-net-worth clients, with a built-in disclaimer.
Needs: Nothing. General information, not tax or legal advice.
Setup
1. Install the skills
Unzip the bundle, or any single skill, into your Claude skills folder. In PowerShell on Windows:
Expand-Archive adi-agrawal-ai-skills.zip "$env:USERPROFILE\.claude\skills" -ForceOn macOS or Linux:
unzip -o adi-agrawal-ai-skills.zip -d ~/.claude/skillsRestart Claude. Then ask for what you need, for example “get US GDP growth from FRED” or “run Brinson-Fachler attribution on this holdings file”.
2. Add the keys
In PowerShell, set both as persistent user variables, then restart Claude and any open terminals:
[Environment]::SetEnvironmentVariable("FRED_API_KEY", "your_fred_api_key", "User")
[Environment]::SetEnvironmentVariable("SEC_USER_AGENT", "Your Name (you@example.com)", "User")A free FRED key comes from the St. Louis Fed. The OECD skill needs no key.
3. Install the Python packages
python -m pip install requests pandas numpy scipy scikit-learn statsmodels matplotlib openpyxl edgartools beautifulsoup4 lxml python-docx4. Add the World Bank connector (optional)
The economic deep research skill also draws on the World Bank. It is a connector, not a skill. In Claude, open Settings, then Connectors, and add a custom connector named World Bank Data360 with this address:
https://maimcpext.worldbank.org/ext/data360/mcpWithout it the skill still runs on FRED and OECD data and notes the gap. Menu labels vary by plan and app version.
Terms
These are my own tools, shared as is, with no warranty. Check the output before you rely on it. They are not investment, legal or tax advice, and the compliance and estate-planning skills do not replace a qualified professional. Data belongs to its providers: FRED, the OECD, the SEC and the World Bank. See Disclosures.