Purpose
Turn a raw exposure dataset across derivatives products into a structured, decision-ready risk summary — computing concentration, counterparty, market, and liquidity metrics, flagging threshold breaches, and ranking the exposures that demand attention first. Solves the problem of risk analysts staring at position blotters and margin files without a consistent framework to surface what actually matters.
When to Use
Invoke this skill when the user asks to:
- Summarize a derivatives book's exposure or risk profile for a daily/weekly report
- Identify where concentration is building across products, underlyings, or maturities
- Assess counterparty and credit exposure across cleared and bilateral positions
- Prepare an exposure briefing for a risk committee or desk head
- Diagnose liquidity risk on a book of options, swaps, futures, or perpetuals
- Flag positions that breach internal risk limits across multiple dimensions
Inputs Required
Before starting, collect:
- Exposure data: Position-level data — product, underlying, notional, direction, maturity (paste, table, or structured list)
- Desk/venue type: Traditional derivatives desk, crypto exchange book, or hybrid — and product mix (futures / options / swaps / perpetuals)
- Counterparty breakdown: Exposure by counterparty or clearing relationship, with credit rating or tier if available
- Risk limits: Internal thresholds — concentration caps, single-counterparty limits, VaR/notional ceilings, margin buffers
- Greeks or sensitivities: Delta, gamma, vega, theta (or net DV01/CS01) if available at book or sub-book level
- Liquidity inputs: Average daily volume, open interest, or bid-ask context per instrument if available
- Reporting horizon: Snapshot (point-in-time) or comparison vs. prior period
If Greeks, liquidity inputs, or limits are missing, compute notional-based proxies and flag each assumption explicitly in the summary.
Instructions
Step 1: Normalize and Reconcile the Exposure Data
Before any metric is computed, standardize the input:
- Convert all notionals to a single reporting currency (state the rate used)
- Net offsetting positions in the same instrument; report both gross and net notional
- Group positions by product type, underlying/asset, maturity bucket (0–7d, 8–30d, 31–90d, 90d+), and counterparty
- Flag any data gaps (missing maturity, missing counterparty, missing sign) — do not silently drop rows
- State the reconciliation result: total gross notional, total net notional, position count
Step 2: Concentration Dimension
Measure how lopsided the book is:
- By underlying: Top 5 underlyings as % of gross notional; flag any single underlying above the concentration cap
- By product: Distribution across futures / options / swaps / perpetuals (% of gross)
- By maturity: Notional per maturity bucket; flag clustering in the 0–7d bucket (rollover/expiry risk)
- By direction: Net long/short skew per underlying — flag one-directional exposure exceeding the cap
- Herfindahl-style read: Note whether exposure is diversified or dominated by 1–2 names, with a one-line interpretation
Step 3: Counterparty Dimension
Measure credit and settlement exposure:
- Top 5 counterparties by net exposure, each as % of total counterparty exposure
- Single-counterparty breach check: Flag any counterparty exceeding its limit, with the amount over
- Cleared vs. bilateral split: % of book facing a CCP vs. bilateral counterparties (bilateral carries higher credit risk)
- Credit-tier weighting: Weight exposure by counterparty tier/rating where available; flag concentration in lower-tier names
- Wrong-way risk note: Flag any case where exposure to a counterparty rises as that counterparty's creditworthiness would deteriorate
Step 4: Market Dimension
Measure sensitivity to price moves:
- Net Greeks (or proxies): Aggregate delta, gamma, vega, theta at book level; per sub-book if data allows
- Directional exposure: Net delta-equivalent notional — the effective long/short the book carries
- Convexity flag: High net gamma or vega clusters that amplify P&L under sharp moves
- Scenario read: Estimated P&L impact under a defined shock (e.g., ±5% / ±10% underlying move) using delta + gamma where available; state the shock assumed
- VaR/notional ceiling check: Compare against the provided limit; flag breach
Step 5: Liquidity Dimension
Measure how hard the book is to exit:
- Days-to-liquidate: Position notional vs. average daily volume per instrument (flag any position exceeding ~20% of ADV)
- Open interest concentration: Position as % of instrument open interest; flag dominant positions that move on exit
- Maturity-liquidity overlap: Cross-reference 0–7d expiry bucket against thin instruments (the dangerous corner)
- Margin/funding buffer: For crypto perpetuals or margined books, note funding-rate exposure and buffer to liquidation thresholds
Step 6: Threshold Breach Roll-Up and Watch List
Consolidate findings:
- List every limit breach across all four dimensions in one table — dimension, metric, limit, actual, severity (Critical / Elevated / Watch)
- Build a prioritized watch list of the top 5–8 exposures ranked by combined severity (a position hitting concentration + liquidity + counterparty limits ranks highest)
- For each watch-list item: state the specific risk and one concrete mitigating action
Output Format
- Structure: Executive summary block, then four dimension sections, then breach roll-up table and watch list
- Length: 800–1,200 words
- Format: Structured document — short prose interpretation per section, metrics in tables, watch list as a ranked numbered list
- Sections: Reconciliation Summary → Concentration → Counterparty → Market → Liquidity → Breach Roll-Up Table → Prioritized Watch List
- Tone: Risk analyst voice — precise, quantitative, flag-first; no hedging on breaches, no narrative filler
Quality Checklist
Before delivering output, verify:
- [ ] Gross notional, net notional, and position count are reconciled and stated up front
- [ ] All four dimensions (concentration, counterparty, market, liquidity) are covered with at least one computed metric each
- [ ] Every threshold breach is captured in the roll-up table with limit, actual value, and amount/severity over
- [ ] Watch list is ranked by combined severity — not listed in input order — with a mitigating action per item
- [ ] Every assumption made to fill a missing input (rate, proxy, shock size) is explicitly labeled as an assumption
- [ ] No instrument with concentration above the cap is omitted from the concentration section
Notes
- The most dangerous exposures usually sit at the intersection of dimensions — a large position that is also illiquid, near-dated, and bilateral. The watch-list ranking exists to surface exactly these compound risks.
- For crypto exchange books, perpetuals introduce funding-rate exposure that traditional frameworks miss — always note funding direction and buffer to liquidation, even when other Greeks are unavailable.
- When Greeks are missing, delta-equivalent notional is an acceptable directional proxy but understates gamma/convexity risk; say so rather than implying full coverage.
- A clean exposure summary is worthless if the underlying data isn't reconciled — never compute metrics on unreconciled rows without flagging the gaps.
How to use this skill
This file is a Claude Code skill in Anthropic's standard format — markdown with YAML frontmatter at the top.
Installing in Claude Code:
~/.claude/skills/<folder-name>/SKILL.md— auto-loads in every session anywhere on your machine (recommended for skills you'll use across multiple projects)<your-project>/.claude/skills/<folder-name>/SKILL.md— auto-loads only when working in that project (recommended for project-specific skills)
Restart Claude Code or open a new session to register.
Using in Claude.ai, ChatGPT, Gemini, or other AI tools:
- Drag this file into a new chat, or copy and paste its contents
- Send your task or request as your next message — the AI will apply the skill where it fits
For Claude Code skill details: https://docs.anthropic.com/en/docs/claude-code/skills