agentleFS
voyager67voyager67/JaneStreet-Agentic-Skills/skills/etf-arbitrage/SKILL.md

etf-arbitrage

Skill13 starsChanged 5 days agoApache-2.0View on GitHub

Analyze ETF primary/secondary-market pricing, basket valuation, creation/redemption economics, premium/discount dynamics, and cross-asset hedging. Use for ETF market making, authorized-participant workflows, basket arbitrage, and ETF liquidity research; not for treating end-of-day NAV as an executable intraday fair value.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 6 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Do not use end-of-day NAV as an intraday executable price.

    line 24
  2. Distinctive:

    Do not assume every market participant can create or redeem directly.

    line 25
  3. Distinctive:

    Do not assume the published portfolio composition file equals the exact executable creation basket in every fund and circumstance.

    line 26
  4. Distinctive:

    Do not infer that a premium/discount is risk-free profit.

    line 27
  5. Distinctive:

    Do not ignore closed local markets, FX hours, holidays, underlying-market halts, or settlement mismatch.

    line 28
  6. Distinctive:

    Do not initiate live ETF, basket, FX, borrow, or derivatives transactions without explicit authorization and the required controls.

    line 29

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

ETF Primary/Secondary-Market Arbitrage

Act as an ETF market-structure and quantitative trading researcher. An ETF is a traded claim on a portfolio with a primary-market create/redeem mechanism. Do not reduce the problem to comparing the ETF last price with stale published NAV.

Workflow

  1. Identify the fund and market structure. Record domicile, listing venue, currency, trading hours, share class, creation-unit size, settlement cycle, authorized-participant access, in-kind/cash/mixed creation, redemption rules, and the authoritative creation/redemption basket source.
  2. Reconstruct the intraday basket value. Price each basket component with synchronized, tradeable reference prices. Apply FX conversion, accrued income, cash components, corporate actions, derivative exposure, fund liabilities/fees where relevant, and basket-specific cash adjustments. Keep official NAV, indicative value, and your own real-time basket estimate separate.
  3. Check basket tradability. For each constituent, measure quote age, spread, depth, market state, trading calendar, borrow availability, price limits, and local-market closure. A quoted basket value is not executable if significant components are stale, halted, closed, or expensive to hedge.
  4. Calculate the deviation consistently. Define premium/discount against the contemporaneous basket estimate using a clear numerator, denominator, currency, and timestamp. Compare ETF bid/ask—not only last trade—to executable basket-side prices. Account for creation/redemption unit sizing and residual inventory.
  5. Build both sides of the economics. For ETF rich to basket: assess ETF sale/shorting, basket purchase, borrow/locate, financing, creation costs, and delivery/settlement. For ETF cheap to basket: assess ETF purchase, basket shorting/borrow, redemption terms, financing, and settlement. Treat these as possible structures, not guaranteed arbitrage trades.
  6. Include frictions and constraints. Model commissions, exchange fees, stamp/transaction taxes, bid/ask, market impact, borrow/recall, financing, FX, creation/redemption fees, taxes, settlement mismatch, minimum size, and capital usage. Include uncertainty in every leg.
  7. Handle asynchronous global markets. For international, fixed-income, commodity, or crypto-underlying ETFs, identify which underlying markets are closed or thin. Estimate fair-value uncertainty from proxy instruments and avoid calling stale underlying prices a reliable NAV anchor.
  8. Evaluate inventory and unwind risk. Track ETF shares, basket constituents, cash, FX, borrow, futures/derivatives hedges, and unsettled creations/redemptions. Stress component gaps, FX jumps, halted constituents, failed creations, settlement fails, and borrow recalls.
  9. Test the mechanism empirically. Study premium/discount persistence, spread/depth, basket liquidity, AP activity if observable, market hours, volatility, and stress periods. Use event-time alignment and avoid using the final NAV as if it were known intraday.
  10. Report actionable thresholds as estimates. Calculate a break-even band and uncertainty interval after all expected costs. State which inputs are tradeable, indicative, stale, model-based, or unavailable.

Guardrails

  • Do not use end-of-day NAV as an intraday executable price.
  • Do not assume every market participant can create or redeem directly; AP eligibility and fund terms matter.
  • Do not assume the published portfolio composition file equals the exact executable creation basket in every fund and circumstance.
  • Do not infer that a premium/discount is risk-free profit; it may compensate for stale underlying marks, liquidity, borrow, financing, taxes, settlement, or inventory risk.
  • Do not ignore closed local markets, FX hours, holidays, underlying-market halts, or settlement mismatch.
  • Do not initiate live ETF, basket, FX, borrow, or derivatives transactions without explicit authorization and the required controls.

Deliverable

Return the fund/venue facts, basket-source and timestamp audit, component tradability table, premium/discount definition, executable two-sided cost model, creation/redemption constraints, residual inventory risks, stress analysis, break-even band, and unresolved data gaps.

Further reading

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.