A user was able to run a swing-low and ascending-higher-low pattern detection using Shibui. And it layered a trend confirmation on top of it (>SMA 200, near 52wk high, above avg. volume).
It's a pretty cool showcase of what Shibui is capable of.
We've now added insider transactions and holdings to https://t.co/g3WzklK6TF, ask Claude: "Find stocks where insiders have filed at least 3 Form 4 transactions in the last 90 days while the stock is down 20% or more from its 52-week high" to get back a list 50 tickers.
most people screen for stock dilution by watching for S-3 / 424B5 filings.
the worst dilution never shows up there.
reverse mergers, PIPEs, convertibles → share count explodes, zero offering filings to flag it.
don't just watch the filings. watch the share count.
@DMN_TDP I submitted the server for inclusion into the official Anthropic directory listing, it's still under review.
What could I improve w.r.t. output transparency?
Found someone using Shibui as ground truth for an MCP tool-use benchmark. Over a thousand queries in our logs. Test questions like "stock whose market cap was closest to X on a specific date" - needs a cross-universe scan across 10K symbols. API wrappers can't do that.
Did some digging on @shibui_finance before connecting it 🔍
✅ Real named operator (German Impressum + VAT ID, not anonymous)
✅ Read-only queries, public market data only
✅ Honest privacy policy — even warns you to keep personal info out of prompts
⚠️ It does log query text + the prompt that triggered it (90 days). So no audits exist yet, and like any MCP connector it could theoretically log inputs or inject via tool results — that's just the architecture, not a knock on them.
Verdict: legit + transparent. Just don't paste account details in those sessions. Solid free stock-data MCP 👌
@DMN_TDP Hey @DMN_TDP , thanks for checking the Shibui Finance MCP server, hope it could deliver the results you were looking for. Happy to answer any questions!
Would the privacy policy prevent you from using Shibui Finance?
@DeepValueReport For LMND the loss ratio is everything. Operating leverage only kicks in if AI underwriting keeps grinding it lower, the number I'd want confirming before adding here.
@qualtrim If FCF outgrew the 16% drawdown, the P/FCF is actually lower now than when he called it cheap at $82. The thesis didn't break, the price did.
@AKWilk The 'not driven by rate increases' part is underrated. Most P&C growth lately is just rate hikes, so mid-teens on actual policy growth is a different animal.
Fox is paying $22B for Roku, 21% above its 52-week high. Roku trades at 4x sales on a 4% net margin.
Profitable media that's cheaper on sales:
Comcast 0.7x, 20% FCF yield
Charter 0.3x, 26%
Disney 1.9x
Sirius XM 1.1x
Fox itself trades at 1.4x sales.
@JohnMuchow CRWV is the more loaded short. Microsoft is ~two-thirds of its revenue, so that 18% is really a bet on one customer more than the AI cycle itself.
@SouthernValue95 The open internet being this small is the real takeaway. The Trade Desk is the only independent at scale and still a rounding error next to any single walled garden.
@linasbeliunas Apple's never done a deal this big though. Beats at $3B is still their largest ever. A Perplexity-sized check just isn't how they operate.
@iancassel 2000 meetings a year is ~8 every business day. At that volume it's not analysis, it's a sourcing operation. The one differential detail pays for the other 1,999.
@BojanRadojici10 The unclaimed early-pay discounts quietly fight that 53.6 DPO. 2/10 net 30 runs ~37% annualized, so stretching those vendors costs more than it saves.
Energy is quietly winning 2026. Best US sector YTD, median stock +29%, 66 of 77 in the green while everyone argued about AI.
Oil just hit a 3-month low on the Iran ceasefire, Strait of Hormuz reopens Friday. The whole sector runs on that one number.
@Greenbackd The passive flow argument is the scariest version. If index buying mechanically pushes capital into the biggest names regardless of valuation, the small-cap premium doesn't mean-revert. It just stays broken.
@DeepValueReport Refinancing risk is the one that kills you quietly. Revenue growing, margins fine, then a maturity hits and the dilution wipes out two years of upside.