- Home
- Skills
- Research & Analysis
- Portfolio CSV Analyst: Allocation, Fees & Gains
Portfolio CSV Analyst: Allocation, Fees & Gains
Turns your brokerage CSV exports into allocation, concentration, fee, gain and income figures, all computed locally. No advice.
Free
Works with the AI Agents you already use
Portfolio CSV Analyst: Allocation, Fees & Gains
Example session with this skill installed
Here are my positions and activity for my two brokerage accounts. The Roth is my IRA, the other is a regular taxable account, and ZNTH and KWIL are individual stocks. Give me the full picture: allocation, how concentrated I am, what I pay in fund fees, and my gains and income for 2026.
- Read your context and instructions
- Compiled the portfolio csv analyst
- Generated the data export
Plan: write funds.csv next to your exports listing ZNTH and KWIL at a 0% expense ratio (as you confirmed), then run analyze.py with your account types and --year 2026.
USD 142,902.35 across 2 accounts and 7 holdings, priced from the export as of 2026-11-06.
| Allocation | Share |
|---|---|
| Equity / fixed income / cash | 86.06% / 10.22% / 3.72% |
| Taxable (Individual ****5678) | 75.66%, USD 108,123.35 |
| Tax-advantaged (Roth IRA ****4321) | 24.34%, USD 34,779.00 |
Concentration [Concentration and the Herfindahl index]: your allocation is 49.78% in one holding (ZNTH), 96.28% in the top five, and the HHI is 0.3185, about 3.1 equally sized holdings.
Costs [Expense ratio]: weighted expense ratio 0.020%, about USD 25.61 a year, covering 89.02% of value. The first run covered only 33.82%, because the export has no ratio for stocks; I did not treat them as 0 until you said they were stocks. QLTR (USD 12,480.00) still has no ratio and stays
unknown, not 0. Its fund's prospectus would fill it.
Gains [Holding period]:
| Amount (USD) | |
|---|---|
| Unrealized total | 49,696.75 |
| Short term / long term | 6,039.00 / 43,657.75 |
| Basis unknown | KWIL, value 7,740.00 |
| Realized 2026 | 1,350.00 (600.00 short, 750.00 long) |
The BRDX lot bought 2025-11-06 is still short term on 2026-11-06: one year exactly is not "more than one year". The 750.00 is
FIFO assumed [Cost basis methods]; the 600.00 uses your broker's basis.
"symbol": "OLDC", "proceeds": "2000.00", "gain": null,
"note": "the export has fewer earlier buys than this sale; the history may start after the purchase"
Income: dividends 312.65, capital gain distributions 88.00, interest 12.34, total 412.99, paid in March, June, September and October.
Cash [Cash share]: 3.72% of the total; 2.97% of the Individual account and 6.04% of the Roth.
Next steps
- The OLDC basis is on your broker's 1099-B; add a history that covers its purchase and I'll rerun.
- Send QLTR's expense ratio and KWIL's cost basis to complete fees and gains.
- This describes your exports; it is not investment, tax or legal advice. A licensed, fee-only fiduciary advisor or a tax professional can help with decisions.
funds.csv
CSV · data export
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
Drop in positions and transaction exports from Fidelity, Schwab, Vanguard, Interactive Brokers, Robinhood or any CSV. A standard-library Python script merges your accounts and reports total value; allocation by asset class, holding, account and taxable vs tax-advantaged; concentration (largest holding, top five, Herfindahl index with its formula); weighted expense ratio and yearly fees in dollars; unrealized gains split short vs long term by lot; realized gains for a year from broker basis or stated FIFO; dividend and interest income by month; and cash share. Missing fund fees show as unknown, never zero, and unpriced positions are listed and left out of the totals. Account numbers are masked. It describes your numbers only and never says what to buy or sell.
No live prices, performance metrics, or wash-sale tracking. Interactive Brokers activity statements yield unknown tax terms. Day-first date formats (DD/MM/YYYY) are not supported.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 9 days ago
- Free with an Agensi account
Requires
Permissions
Runs the bundled scripts locally on the files the user names. Writes output only inside the user's project, after asking, and makes no network requests.