AILC-19 · Drawer: Investing & markets
/portfolio-checkup
Checks your portfolio against your own rules: hidden overlap between funds, fees in euros, concentration, currency.
Ask it: “Do my ETFs overlap? Check my portfolio.”
When it runs
Review a personal investment portfolio (a broker export, a spreadsheet or a typed list) against the investor's own stated goals and rules: concentration, overlap between funds, costs, currency and country exposure, cash drag, and whether the mix still matches their time horizon. Returns a plain-language health check with numbers, not buy or sell calls. Use this skill whenever the user shares holdings and asks to "check my portfolio", "am I diversified", "how much am I paying in fees", "do my ETFs overlap", "rebalance check", "too much in one stock", or before adding new money. Do NOT use to pick stocks, time the market, recommend specific products, give tax advice, or for someone else's portfolio without their consent.
The full file
--- name: portfolio-checkup description: >- Review a personal investment portfolio (a broker export, a spreadsheet or a typed list) against the investor's own stated goals and rules: concentration, overlap between funds, costs, currency and country exposure, cash drag, and whether the mix still matches their time horizon. Returns a plain-language health check with numbers, not buy or sell calls. Use this skill whenever the user shares holdings and asks to "check my portfolio", "am I diversified", "how much am I paying in fees", "do my ETFs overlap", "rebalance check", "too much in one stock", or before adding new money. Do NOT use to pick stocks, time the market, recommend specific products, give tax advice, or for someone else's portfolio without their consent. --- # Portfolio checkup Most personal portfolios fail quietly, not dramatically. Three "different" global ETFs that hold the same 500 companies. A favourite stock that grew into 40% of the money without anyone deciding it should. Fees of 1.5% a year that look small and cost a third of the gains over twenty years. Cash waiting "for a better moment" for three years. None of these show up when you look at the total value going up. They show up when someone counts. This skill counts, then compares what it finds with the rules the investor set for themselves. It does not tell anyone what to buy. It shows them where their portfolio has drifted from their own plan. ## Workflow ### 1. Get the holdings and the rules Holdings: name, ticker or ISIN, amount or value, currency. A broker export is ideal; a typed list works. Rules, asked in one round (numbered): 1. What is this money for, and when will you need it? 2. Your target mix, if you have one (e.g. 80% shares, 20% bonds or cash). 3. The most you want in any single company, and in any single country. 4. How would you feel if it fell 30% in a year? (sold / waited / bought more) 5. Anything you will not hold, for any reason. If the user has no rules, say that this is the first finding, and help them write three before checking anything else. ### 2. Run the nine checks 1. **Mix vs target:** shares, bonds, cash, other, as % of total. 2. **Single-name concentration:** largest holdings as % of total, including exposure through funds where the data is available. 3. **Fund overlap:** for ETFs and funds, the share of top holdings they have in common. Name the overlapping companies. 4. **Country and currency:** exposure by country and by currency versus where the investor lives and spends. 5. **Sector tilt:** largest sectors vs a broad world index. 6. **Costs:** ongoing charge per fund, the weighted average, and the cost in money per year. Show the 20-year effect at an assumed return, with the assumption stated. 7. **Cash drag:** cash as % and how long it has been there, if known. 8. **Horizon fit:** money needed within 3 years sitting in volatile assets. 9. **Rule breaks:** every place the portfolio breaks the user's own rules. Use current data where available, with the as-of date. Where you cannot see a fund's holdings or fees, say "not checked" and how to find it (the fund's factsheet or KID). ### 3. Report ``` ## Portfolio checkup, [date], total [value, currency] **In one line:** [the most important finding, with its number] | Check | Finding | Your rule | Status | |---|---|---|---| | Largest single company | Apple 14% (direct 6% + via 3 ETFs 8%) | max 10% | BREAK | **Three things worth a look** (ranked by money at stake) 1. [Finding, number, why it matters for THEIR goal] **What looks fine:** [checks that passed] **Not checked:** [and how to check] Education, not financial advice. Decisions are yours. ``` ## Judgment rules **Their rules, not yours.** A 100% equity portfolio is not "too risky" for someone with a 30-year horizon who said they would hold through a crash. Flag only conflicts with their stated goals and with plain arithmetic. **Show money, not just percent.** "0.9% in fees" lands harder as "€1,350 this year, on €150,000". **No product picks.** Never name a fund or stock to buy or sell. If asked, explain which kind of question to take to a licensed adviser. **Overlap is the hidden one.** Always look through funds. It is the check people skip and the one that most often changes the picture. ## Example A user holds an MSCI World ETF, an S&P 500 ETF, a tech ETF and Apple shares, and believes they are diversified across four holdings. The look-through shows the same seven US companies make up 31% of the total, Apple alone 11% against the user's own 10% limit, and 71% of the money is in US dollars while they spend in euros. Nothing was bought by mistake; it just added up.