MIS v2 — Where We Are (plain-English map)

Updated: 2026-05-27 ~3 PM. One page so the spinning stops. Re-open this anytime you lose the thread.

Old system vs new system (this is probably what's spinning)

OLD MIS = your original MIS sheet + script. It is still running, untouched — it's the thing that emails you the morning + 2:05 PM briefs. We are not building on it; we left it alone so nothing breaks. (If a morning email goes missing, that's an OLD-MIS trigger, separate from the new build.)
NEW MIS v2 = a brand-new, separate sheet + brain we've been building. It does not email you. I run it on demand. This is where all the new work lives (the entry tickets, the holdings, the cross-check).
They do not touch each other. Nothing is tangled. When v2 is ready, the old one gets retired — not before.

Where the information comes from (only 3 sources — that's it)

Google Finance — live prices, % moves, momentum (3/10/30-day), 52-week range, the indexes (S&P/Dow/Nasdaq/VIX/10Y). ~15-minute delay. (Tagged LIVE on the ticket.)
Finnhub — next earnings date + market cap. Pulled on a schedule and cached (stored), with the "last pulled" time shown. (Tagged CACHED.)
Your broker emails — Fidelity "Position Summary" = your holdings; Fidelity/Schwab execution emails = your trades. Schwab's weekly email also carries positions. This is how it knows your positions without you logging into Fidelity. (Tagged from the email + its date.)

The verdict/stop/target/sizing are then computed by the engine from #1. Every number on the ticket says which of these it came from + when (the DATA FRESHNESS box). If something isn't pulling right, that box turns red so you know not to trust it.

Do you "run reports"? — No, that's the point

The goal is always-there, glance-don't-run: you open the portal (or text the bot) and it's current. You don't generate a report; you look. Today I'm the one running it on demand — the auto/always-on part is what we're building toward.

Is your information "out there" / at risk? — No

The workspace + all the new code = a private GitHub repo (only you).
Your memory/notes = local on your machine only — never uploaded, Mildred never sees it.
The portal (ops.hookstreetservices.com) sits behind Cloudflare Access — only your Google login opens it. Not public.
No account numbers and no API keys are in the code. Keys live in the sheet's hidden Script Properties, never in git.
Worst case if anything leaked: a link, not your money or your logins.

Where we actually are right now

✅ The brain works and is proven on 2 tickers (INTC = ADD, NVDA = REJECTED) — no contradictions.
✅ All your holdings loaded — $32,745 across all 5 of your accounts (Fidelity Margin 0600 + Cash 5692, Schwab ZT 898 + SWPF 436 + TV 241); kids' accounts (5378, 4073, $1,751) separate. ORCL close applied. (Fidelity from email, Schwab from your screenshots.)
✅ Test Entry finalized for use — reads like a brief, shows upside % + risk, is tweakable (budget / hold-years / max-loss cap → re-sizes + dual stop), behavioral gut-check, plain how-to, scenarios, smart data-freshness, and a TRACK RECORD (learning loop — what MIS said before + how it did, accumulates daily).
✅ VISUAL DASHBOARD shipped (fn=dashboard, auto-published to the portal): market row (S&P/Dow/Nasdaq/VIX/10Y/Gold) + per-ticker cards with live ~130-day price charts + 52w range bars. The "stop looking like a briefing" answer.
⏳ Not yet (captured, none half-broken): real performance / margin levered-unlevered (needs your Fidelity Positions CSV — cost basis) · Schwab auto (wire its weekly email) · Telegram → MIS v2 (Phase 2) · charts for sectors/ETFs · self-learning auto-accumulation (one trigger-scope enable) · the full universe.

Lessons locked today (2026-05-27 PM)

Advisory, never autonomous + verdict depends on intent/horizon (swing REJECT can be a fine 5-yr core) — feedback_mis_advisory_not_autonomous.
Charts, not prose is the fix for "feels like a briefing"; charts = market/tickers/sectors, not holdings.
No false numbers: only show cost-basis/P&L where the data is current; the repo CSVs are stale (March/May 1) → need a fresh Positions export.
Discipline confirmed: stay at 2 tickers until the brain + surfaces are perfect; Sam explicitly approved the pace.

The honest bottom line

Two systems, three data sources, your stuff is private. The brain is real at 2 tickers; all holdings consolidated; the Test Entry is usable and the visual dashboard is live on the portal. Next: performance/margin (your CSV), Telegram wiring, sector charts — then scale. Nothing wasted, nothing exposed.

Map · updated 2026-05-27 ~4:45 PM · Claude Code.