1 — Portfolio X-ray dashboard (visual + actionable) #
Connect to my IBKR account and build me an actionable portfolio X-ray as a
self-contained HTML dashboard I can open in my browser.
GOLDEN RULE — be GRAPHICAL-FIRST: show every piece of information as a visual
(chart, gauge, heatmap, treemap, colored badge, progress bar, or sparkline)
wherever it's possible, with the numbers as labels inside the visual. Use plain
text only for short titles and the improvement suggestions. If you catch yourself
about to write a paragraph or a plain table, convert it into a chart instead.
Use color consistently: green = good/up, red = risk/down.
Pull my live positions, account value, cash/margin used, and YTD + 1-year
performance, then render:
- KPI tiles: net liquidation value, YTD + 1-year returns (with up/down arrows and a
sparkline), and a margin/leverage GAUGE
- equity curve: line/area chart for the year
- sector allocation: donut chart
- top holdings: horizontal bar chart with % weight labels
- winners & losers: one diverging bar chart (green right / red left)
- LOOK-THROUGH exposure: a bar chart or treemap of my TRUE exposure to each major
company — DIRECT holdings PLUS INDIRECT holdings inside my ETFs (e.g. real NVDA /
AAPL / MSFT once you look inside QQQ and SCHG). Use the ETFs' known top holdings
and weights; label it as an estimate.
- concentration: a treemap or heatmap so over-weight positions/sectors pop visually
- risk flags: colored badges / traffic-light indicators (single-stock, sector,
look-through concentration, margin/leverage) — NOT a paragraph
The only text section: 3-5 actionable improvement suggestions, each as a compact
card with a severity color and one line of reasoning.
Where IBKR doesn't provide the data (like exact ETF holdings), still visualize it
with best estimates and mark it as estimated — don't drop it. Output as one
self-contained HTML file.
2 — Find a new stock, then deep-dive it (hunt → analyst report) #
Help me find a new stock to buy that I don't already own, then deep-dive it like an
analyst.
First, HUNT — but before digging in, ask me how I want to hunt: (a) pick from a few
investment themes you propose, (b) I give you a specific style or sector, or (c) if
I'm not sure, just default to quality businesses that have recently sold off and look
undervalued. Wait for my pick (unless I've already told you). Then, using my choice,
look through IBKR's investment themes plus a couple of names that are beaten down or
unusually in the news right now. Verify they're real tickers, skip anything I already
hold, and give me a quick ranked shortlist of ~10 with a one-line reason each.
Then DEEP-DIVE your top 3 picks — give me a full equity research report for each,
every claim evidence-based, clearly separating facts from inference from speculation:
1. Business model — how it makes money, revenue segments, is it sustainable
2. Financial health — operating cash flow, free cash flow, debt
3. Valuation — multiples vs peers (P/E, EV/Revenue, P/B) AND a rough DCF /
intrinsic-value estimate (show the key assumptions and give a fair-value RANGE,
not one magic number); is it cheap or expensive vs its growth?
4. Macro & competitive — rates, regulation, competitors; the biggest threat
5. Catalysts — both the obvious ones and the hidden/underappreciated ones
6. Risks — regulatory, concentration, competitive, execution, valuation
7. Investment thesis — strongest bull AND bear case, then: is it asymmetric
risk/reward at today's price, and what are the key metrics to watch?
Give me everything as a visual HTML report I can open in my browser. Cite your
sources and don't make up tickers or numbers.
3 — Where's my stock headed next year? (macro + fundamentals + technicals + catalysts) #
Connect to my IBKR account. For each of my main holdings (or a stock I name), give
me a grounded view of where it might go over the next year — combining macro,
fundamentals, technicals, and upcoming catalysts. I know it's not a prediction; I
want a reasoned range, not a guess.
Handle ETFs differently from individual stocks: for an ETF (like QQQ or SCHG), base
the outlook on the index it tracks — the macro/rate backdrop, the index's overall
valuation and earnings growth, and how concentrated it is — not company-level
fundamentals.
For each holding:
- Technicals — where it's trading vs its 52-week range, the trend, and momentum.
- Fundamentals — growth, valuation, and whether the business is improving or
deteriorating.
- Macro — the rate / sector / demand backdrop that helps or hurts it.
- Upcoming catalysts — earnings dates, product launches, or events in the next 12
months that could move it.
- Then give me bull / base / bear scenarios with rough 1-year price targets and the
reasoning for each, plus a realistic range using its historical volatility as a
sanity check.
Show it as a visual HTML report — a fan chart or bull/base/bear bands per stock,
with the catalysts marked on a timeline. Be clear what's fact vs your judgment, cite
sources for fundamentals and catalysts, and don't fabricate any numbers or earnings
dates.
4 — Daily portfolio brief + buy-alerts (scheduled) #
Every day, build me a portfolio brief and email it to me as a nicely formatted,
visual HTML email (use my Resend API key). Connect to my IBKR account and cover:
1. Snapshot — total portfolio value and the day's change (color-coded green/red),
plus my biggest movers with one line on what drove each.
2. News & events — overnight or market news affecting my holdings, and a
macro/catalyst watch (rate decisions, big economic data, sector events) that
could move my positions.
3. Upcoming earnings — any of my holdings reporting soon, with the date and the
expected move.
4. Buy alerts — flag any current holding that looks like a good buy to add now
(within ~10% of its 52-week low, OR down 20%+ from its high, OR an elevated
implied-volatility percentile / fear); give the price, the trigger, and one line.
5. Act today — a short list of anything I should actually do.
Make it scannable and visual: color for up/down, simple bar/sparkline cues where
they help, clear sections. (Email can't run interactive charts, so keep visuals as
styled HTML and static cues.) Also generate the full interactive HTML portfolio
dashboard (my visual X-ray) and ATTACH it to the email — so the body is the quick
daily read and the attachment is the full visual I can open in my browser. If a
section has nothing notable, say so in one line. Cite sources for news and earnings
dates, and don't fabricate them.
5 — Options position manager (roll · close at 50% · assignment watch) #
Connect to my IBKR account and act as my options position manager — I sell premium
(covered calls and cash-secured puts), so go through every open option I have and
tell me what to actually do with each one today.
For each open option, first lay out the facts:
- Days to expiry, the strike, and whether it's a covered call or a cash-secured put
(flag any call that isn't covered by shares, or any put I don't have the cash for).
- Where the stock is now vs my strike and my breakeven — in the money or out, and by
how much.
- Approximate assignment risk: the connector gives implied volatility but NOT the
Greeks, so compute the delta yourself (Black-Scholes from IV, spot, strike, DTE)
and read it as a rough assignment probability. If option_midpoint_iv looks invalid,
fall back to implied_vol, and say when the data is thin.
- Profit so far as a % of the premium I sold it for, and the current P/L.
Then — and this is the part that should drive the decision, not just the delta and a
50% rule — read what's actually happening with the underlying between now and expiry:
- Technicals: the trend, and where the stock is sitting versus the support/resistance
that matters for my strike. Is it about to break through my strike, or stalling
short of it?
- Catalysts before expiry: ANY earnings date inside the life of the option is the big
one — assignment risk and IV both jump around earnings — plus product launches or
events. And the macro backdrop (a Fed decision, CPI, a sector event) that could
move it before it expires.
- The stock itself: is my original reason still intact — for a put, would I still be
happy to own it at the strike; for a call, do I actually want to keep the shares or
am I fine letting them go?
Now give me the call for each one, justified by that read (not just the mechanics) in
a line or two:
- HOLD — thesis and technicals are fine and there's still premium to collect; let it
ride.
- CLOSE — I've captured most of the premium (e.g. ~50%+) and the risk/reward of
holding the rest is poor, OR a catalyst before expiry isn't worth sitting through;
take it off.
- ROLL — under pressure or near expiry and I want to stay in: suggest a specific roll
(out in time, and down/up in strike, around the support/resistance level) with the
target strike, expiry, and the rough net credit/debit — and roll PAST the next
earnings date, not into it.
- ASSIGNMENT RISK — deep in the money or about to be assigned: spell out what
assignment would mean (shares called away, or stock put to me at the strike), and
whether — given the technicals and the catalyst calendar — I should let it happen or
defend it.
Build it as a self-contained, GRAPHICAL-FIRST HTML report I can open in my browser —
show every position as its own card, sorted most-urgent first (nearest expiry +
highest assignment risk at the top). On each card: a DTE countdown, an
assignment-risk GAUGE (the delta, green→red), a moneyness bar (spot vs strike vs
breakeven), a premium-captured progress bar (green), a one-line technical read, a
CATALYST flag if anything lands before expiry (e.g. a red "⚠ earnings in 5 days"
badge), and a big colored ACTION badge (hold = green, close = blue, roll = amber,
assignment risk = red). Up top, KPI tiles: number of open positions, how many need
action now, how many have a catalyst before expiry, total premium still at risk, and
premium captured so far. End with a "what I'd do today" list, ordered by urgency.
Be clear which numbers are live from IBKR (price, IV, strikes, P/L, the technicals)
vs computed (the delta / assignment odds) vs from the web (earnings dates, macro
events) vs your own judgment (the hold/close/roll call). Cite sources for earnings
dates and news, don't fabricate any dates, flag any missing or unreliable option data
instead of guessing, and treat this as analysis, not advice.
6 — Options opportunity spotter (is now a good time to sell a CC / CSP?) #
Connect to my IBKR account and help me spot good options-selling opportunities —
covered calls and cash-secured puts — judged on whether NOW is actually a good time
to sell premium (technicals, implied volatility, and the macro backdrop), not just a
list of strikes.
First, before you run anything, ask me up front if there are any specific tickers I'm
eyeing — and let me skip if I don't have any. Then run two sweeps end to end without
stopping to ask again:
1. My whole portfolio — for each holding where I own at least 100 shares, check if
it's a good moment to sell a COVERED CALL (e.g. stock near resistance / overbought
/ rich IV). And for names I hold or have cash to add to, whether it's a good moment
to sell a CASH-SECURED PUT (e.g. stock near support / pulled back / elevated IV).
Surface only the setups that genuinely look good right now — don't force trades.
2. Any tickers I named up front — judge whether now's a good time to sell a CSP or CC
on each.
For every opportunity, give me: WHY now is a good time (the technical + IV + macro
read in a line or two), then the trade — strike, expiry, premium, IV, delta, and
annualized yield. Target ~0.20-0.30 delta and 30-45 days to expiry, skip expiries
that span an earnings date, and only liquid contracts (tight bid/ask, decent open
interest). For puts, check I have the buying power.
Build it as a self-contained, GRAPHICAL-FIRST HTML report (color-coded opportunity
cards, IV-percentile and annualized-yield bars, a quick technical read per name). The
connector gives IV but not the Greeks, so compute delta yourself. Be clear which
numbers are live from IBKR (price, IV, technicals) vs your own read (the macro view
and any earnings dates) — cite sources for those, and don't fabricate earnings dates.
Flag any missing or unreliable option data, and treat this as analysis, not advice.
source & further reading
gist.github.com — original article
apple-container-cli-skill.md
Module 2 (Vector Search) — FAQ Additions
The spelled-out intro to neural networks and backpropagation: building micrograd.ipynb