GS

    GPW Stock Research - 4 Value Investing Masters

    1

    Analyze Warsaw Stock Exchange stocks through Buffett, Munger, Duan Yongping and Li Lu lenses - ending in PLN buy/sell ranges.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more
    GS

    GPW Stock Research - 4 Value Investing Masters

    Example session with this skill installed

    Analyze XTB using the four masters team framework

    • Read your context and instructions
    • Compiled the gpw stock research

    XTB S.A. (GPW: XTB) — four masters team analysis

    Date: 2026-07-07

    Price: 122.58 PLN (52-week high)

    Market cap: 14.41 bn PLN (117,569,251 × 122.58, computed by hand)

    1. One-sentence conclusion

    XTB is a superbly run, highly profitable machine for collecting spreads from retail CFD speculators (~90% of revenue) with a real but still marginal investing pivot — a good company at the wrong price at the wrong time: at 122.58 PLN you buy a cyclical, regulation-exposed business at a growth multiple, at a 52-week high, with no margin of safety.

    2. Ratings

    LensMasterRatingKey judgment
    Business essenceDuan Yongping★★★☆☆Two businesses in one app: a high-margin "toll" on CFD speculators (94% of revenue, ~72–80% of clients lose money — XTB KID ~74%) plus a small, low-margin investing pivot (stocks/ETF/IKE ~4%). Excellently run, but the core is ethically borderline and unpredictable.
    Financials & valuationBuffett★★☆☆☆Quality ★3 (ROE 25–53%, ~2.6 bn PLN net cash, no debt), valuation ★2: base value ~88 PLN, price 122.58 = ~40% premium, zero margin of safety.
    Risk / inversionMunger★★☆☆☆22× P/E vs peers (Plus500 ~11–15×, IG ~10–17×) for a business with HIGHER earnings volatility and regulatory concentration. Fortress balance sheet rules out ★1, but the risk of lasting disappointment is high.
    10-year certaintyLi Lu★★☆☆☆Profit engine (CFD ~90%) runs against ESMA/KNF regulation; pivot credible but unproven in the P&L. 10-year test = "don't know" = don't buy. Management ★3.

    Composite: 2.3 / 5 — high operating quality, wrong price, high cyclical-regulatory risk, low 10-year certainty.

    3. Tensions between the masters

    Quality vs cyclicality. Buffett sees an excellent business bought at a bad price — the defect is in the price, not the company. Munger flips it: XTB is a cyclical market-maker, and London prices Plus500/IG at ~11–15× precisely because they are cyclical and regulation-capped. So XTB isn't cheap relative to growth, it's

    expensive relative to its true peer group — multiple compression to peers alone means −35% to −50%.

    The profit engine. Earnings are real in cash but extremely unrepeatable (10× quarterly swing: Q3'25 53.2 m → Q1'26 ~535 m PLN). 88.5% of the record Q1'26 came from commodity volatility (gold/silver) — a windfall, not a new base. All four agree: anchoring on TTM 985 m PLN is a mistake; calendar 2025 profit was

    643.8 m PLN (−24.9% y/y).

    Full agreement: bulletproof balance sheet; no margin of safety at 122.58; "I earn when the client loses" is a structural regulatory risk (20 m PLN KNF fine = start of a trend); the founder keeps selling (43.78% → 35.78%).

    4. Tiered recommendation

    Base intrinsic value (normalized profit ~800–900 m PLN through the cycle):

    ~88 PLN/share. Bear ~45–46 PLN, bull ~140–157 PLN.

    StrategyAt 122.58 PLNBuy belowHoldSell above
    Aggressive (believes in lasting 30%+ ROE)HOLD (near sell zone)88 PLN88–140140 PLN
    Moderate (base + ~20% margin)OVERVALUED / trim72 PLN72–118118 PLN
    Conservative (weights cyclicality + CFD/ESMA risk)AVOID / sell60 PLN60–9595 PLN

    Net: no buying opportunity. Entering today means paying for the peak of the volatility cycle.

    5. Thesis falsification conditions

    The bull case breaks if any of these happen:

    1. ESMA/KNF cuts leverage on commodity CFDs (now 10:1) or bans the product class.
    2. Another year of net margin <30% on rising revenue (2025 costs +48% vs revenue +14.6%).
    3. Stocks/ETF + interest income stay below ~10% of revenue for 2–3 years.
    4. 2–3 more flat quarters like Q3'25 (~50 m PLN profit).
    5. The founder drops below the declared ~30% stake or speeds up selling.

    6. Mirror test (thesis in 5 sentences)

    XTB earns ~90% from spreads on retail CFD speculators, most of whom lose money — highly profitable but extremely dependent on market volatility and under lasting ESMA/KNF pressure. It is genuinely pivoting toward long-term investing (2.16 m clients, +64% y/y), but that segment is <4% of revenue and structurally low-margin. At 122.58 PLN the market prices XTB as a growth compounder (22× 2025 earnings, 5.67× book), paying a 50–100% premium to similar peers by extrapolating a commodity windfall quarter. Base value is ~88 PLN, so there's no margin of safety, and the founder is selling.

    Watchlist, not a buy at 122 PLN.


    Next steps

    • Saved: reports/XTB/XTB-analysis-20260707.md + thesis tracker XTB-thesis.md
    • Run /gpw-thesis XTB after each quarterly report to check falsification conditions
    • Price alert: re-run /gpw-analyze XTB if the price drops below 95 PLN

    Research tool, not investment advice. All ratios computed with financial_rigor.py, not mental math.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Turn your agent into a disciplined equity analyst for the Warsaw Stock Exchange.

    14 research playbooks built on four value-investing lenses — Duan Yongping (is this a good business?), Buffett (numbers and valuation), Munger (how this company dies), Li Lu (the 10-year test) — with hard anti-bias rules:

    • every key number from 2 independent sources, market cap recomputed by hand
    • exact decimal arithmetic via bundled Python tools, no mental math
    • missing data is "NO DATA", never a guess
    • every verdict ends in tiered PLN ranges (aggressive / moderate / conservative) — no fence-sitting

    What you can ask:

    • "analyze XTB" — full four-lens report with buy/sell ranges (10–20 min)
    • "scan the watchlist for buy signals" — momentum breakouts confirmed by fundamentals
    • "review my portfolio" — SELL / BUY / WAIT orders with price levels
    • "does my thesis on CDR still hold?", "why did KGH drop today?", "read PKN's quarterly results"

    Research tool, not investment advice. Open source (MIT): github.com/hkn-33/gpw-research-workshop

    GPW only. Fundamentals scraped from public Polish sites - quality depends on their coverage. Small caps get "researchability C" (first-principles mode). Not investment advice; executes no trades

    How to use

    1. 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. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 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 23 days ago

    • Free with an Agensi account

    Requires

    Yahoo
    python3

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Network Access

    Allowed Hosts

    query1.finance.yahoo.com

    Shell runs bundled Python tools. Network: public quotes from Yahoo Finance + web pages of public financial sites. Writes reports to reports/ and reads your portfolio.md. No API keys, no credentials.

    Listed23 days ago

    What's inside

    Frequently Asked Questions

    Any agent that supports SKILL.md and can run Python and browse the web: Claude Code (best - the team analysis runs 4 subagents in parallel), Codex CLI, Cursor, Gemini CLI. Agents without subagents run the four lenses in sequence.

    Quotes and price history come from Yahoo Finance’s public API via the bundled Python tools. Fundamentals, ESPI filings and news come from public sites (biznesradar.pl, stooq.pl, bankier.pl) that the agent reads through web search. Decision-critical numbers are checked in 2 sources, and all arithmetic uses exact decimals. Quotes are delayed, not real-time.

    4 research playbooks, 7 agent roles (scout, 4 master lenses, auditor, timing), 5 stdlib-only Python tools, the methodology doc, a GPW watchlist and a portfolio template.

    Any GPW ticker that Yahoo Finance covers. The scanner watchlist includes WIG20, mWIG40 and selected sWIG80 names. Thinly covered small caps get “researchability C” and are analyzed from first principles.

    They are flagged as a separate, correlated political-risk class, and the Munger lens explicitly covers state and regulatory risk. There is no automatic numeric risk premium.

    No. It’s a research tool: it executes no trades, and decisions are yours.