A Skill is a folder containing a SKILL.md file — plain Markdown with a short YAML header. The agent reads the header of every installed Skill; when your question matches what the header describes, it loads the rest of the file and follows the instructions inside. There is no code to compile and nothing to run. A Skill bundles three things: a tool manual (endpoints, parameters, auth), a translation layer from plain English to the right API call, and rules about how the agent must report the answer.

Do Quant Data Skills only work with Claude?

No. The SKILL.md format is a Claude convention, but the body is just instructions plus an HTTP contract. Copy everything below the YAML header and paste it into any agent system prompt: ChatGPT Custom GPTs, Cursor and Windsurf rules, Cline, Gemini CLI, GitHub Copilot instructions, n8n, Dify, CrewAI, AutoGen, LangChain, or code written directly against the OpenAI or Anthropic SDKs. If your tool can send an HTTP request with one custom header, it is supported.

Where do I put my Quant Data API key?

In the environment variable QUANTDATA_API_KEY, or in whatever secret store your agent platform provides. Never write it into SKILL.md — that file is meant to be copied, committed and shared, so a key inside it will leak. The max-pain skill needs no key because it reads free public pages.

What stops an AI agent from turning this into trading advice?

Each SKILL.md contains explicit rules: never present output as a buy or sell recommendation, never give personalized investment advice, always carry the disclaimer, always state the baseline next to any accuracy figure, and always flag when a statistic is being applied to an asset it was never validated on. Those rules are the reason the Skills exist rather than a bare API reference. Nothing here forecasts price.

AI Skills

Three installable Skills that give an AI agent Quant Data's market analytics — and, just as importantly, rules about how to report them honestly. Free to install. Two of them need an API key; one does not.

What is a Skill, if you have never used one?

A Skill is a folder with a SKILL.md file in it. SKILL.md is plain Markdown with a small YAML header. That is the whole format — there is no code to compile and nothing to run.

Your agent reads the header of every installed Skill. When you ask a question that matches what the header describes, the agent loads the rest of that file into its context and follows the instructions inside. It is closer to handing someone a one-page briefing than to installing software.

So a Skill is three things at once:

A tool manual

Base URL, endpoints, query parameters, the auth header, the error table. The mechanical part an agent would otherwise have to guess at.

A translation layer

You say "what kind of day is NVDA having". The Skill tells the agent that means GET /v1/read/NVDA — and that "gold" is not a ticker, but GLD is.

Rules about the answer

The part people skip, and the part that matters. No buy/sell framing, the baseline stated next to every accuracy number, out-of-distribution use flagged out loud.

You can read every line of these files. Nothing is hidden, and nothing leaves your machine except the one HTTPS request you asked for.

The three Skills

Skill	What your agent gains	Key
`quantdata-daily-bias`	Day-type probabilities for the session in progress, the most similar historical days, and a fact-checked database of price-action claims	$30/mo
`quantdata-weis-wave`	Volume-price wave structure and the five Weis events, each carrying its measured win rate and sample size	$30/mo
`quantdata-max-pain`	Options max pain, open-interest call and put walls, put/call ratio, gamma exposure	free

Both paid APIs together are $50/mo. Max Pain is free and always will be — it reads public pages on this site and never sees a key.

quantdata-daily-bias

Teaches the agent to call /v1/read, /v1/match and /v1/claims on api.quantdata.uk, and to report the result as a five-way probability distribution rather than a label. Triggers on questions like "what kind of day is TSLA having", "is this a trend day", "which historical days looked like this", or "is it true that 80% of breakouts fail".

The instruction block covers symbol resolution (US tickers, digits.HK, 6-digit A-shares, =F futures, the GOLD-is-Barrick trap), the interpretation rules — always state the 37% majority baseline next to the 66% accuracy, always give top-2 coverage, always flag that single stocks are out of distribution — and the full error table.

About the Daily Bias API · Learn day types first

quantdata-weis-wave

Teaches the agent to call /v1/weis and to read effort against result: the wave in progress and its volume relative to the last twenty, the eight most recent completed waves, and any of the five classical events that have fired.

Critically, it carries the measured scoreboard. Two of the five events confirm the tradition; two measured reversed. An agent without this Skill will read "no demand" as bearish because every textbook says so. An agent with it reports that the event fired, that the tradition calls it bearish, and that across 22,262 measured instances it resolved bearishly only 45.6% of the time.

About the Weis Wave API · Learn volume-price analysis first

quantdata-max-pain

Needs no key and no account. It points the agent at the free public pages at quantdata.uk/max-pain/ and teaches it what the numbers mean — open interest, max pain, call and put walls, put/call ratio, gamma exposure and the gamma flip level.

It also teaches the agent to state the pinning theory as a theory rather than a fact, to always report the snapshot date because open interest is end-of-day data, and to describe gamma numbers as model estimates that rest on unobservable assumptions about who is on which side of each contract.

Browse the free max-pain pages

Install (Claude Code, Claude Desktop, the Claude apps)

Skills live in a skills directory. Personal skills are available in every project; project skills are checked into a repository and shared with everyone working in it.

Scope	Path
Personal	`~/.claude/skills/<skill-name>/SKILL.md`
Project	`<your-project>/.claude/skills/<skill-name>/SKILL.md`

# personal — available in every project
mkdir -p ~/.claude/skills
cp -R quantdata-daily-bias ~/.claude/skills/
cp -R quantdata-weis-wave  ~/.claude/skills/
cp -R quantdata-max-pain   ~/.claude/skills/

# or project-scoped — committed alongside your code
mkdir -p .claude/skills
cp -R quantdata-max-pain .claude/skills/

Restart the agent, or start a new session, so it picks up the new folder. Then ask a normal question — you never invoke a Skill by name. To check it landed, ask "which Quant Data skills do you have?"

Set your key

Two of the three need one. Never paste a key into SKILL.md — that file is meant to be copied, committed and shared, so a key inside it will leak. Put it in your environment instead:

# ~/.zshrc on macOS, ~/.bashrc on most Linux
export QUANTDATA_API_KEY="qd_your_key_here"

Keys start with qd_. For an agent with no shell — Claude Desktop, a Custom GPT, an n8n node — use that platform's secret store instead. Full walkthrough on the AI agent setup page.

Using these without Claude

Nothing here is vendor-specific. SKILL.md is a Claude convention, but the body is just instructions plus an HTTP contract. Copy everything below the YAML header and paste it into any agent's system prompt, custom instructions or persona field, then point that agent at wherever your platform keeps secrets.

Claude Code Claude Desktop Cursor / Windsurf / Cline ChatGPT & Custom GPTs Gemini CLI n8n / Dify / CrewAI Anything that sends a header

The raw content

Below is the paste-ready version — the same instruction set the three SKILL.md bodies carry, condensed into one system prompt. Fill in your key and drop it into any agent. The full files live in the skills/ folder of the Quant Data distribution as skills/quantdata-daily-bias/SKILL.md and siblings.

System prompt — all Quant Data endpoints (copy this)

You have access to the Quant Data APIs.

Base URL: https://api.quantdata.uk
Header on every request:  X-API-Key: <your qd_ key>

ENDPOINTS
  GET /v1/read/{symbol}       Day-type probabilities for the current session
                              (TrendUp / TrendDown / Range / TrendingRange / Reversal).
                              Optional ?session=asia — 24h futures only, unvalidated.
  GET /v1/match/{symbol}?k=3  Most similar historical sessions (date, similarity, day_type)
                              from a 2010-2026 S&P 500 E-mini intraday library.
  GET /v1/claims              Fact-checked price-action probability claims, measured against
                              3,946 real trading days.
  GET /v1/weis/{symbol}       Weis Wave volume structure: the wave in progress, the last
                              8 completed waves, and detected events with measured win rates.

Free, no key, ordinary GET on the public site:
  https://quantdata.uk/max-pain/{ticker}   Max pain, OI walls, put/call ratio, gamma exposure.
                                           Lowercase the ticker in this URL: NVDA -> /max-pain/nvda.

SYMBOLS
  US tickers as-is (NVDA) for the API; lowercase them in max-pain page URLs.
  Hong Kong = digits.HK (Meituan = 3690.HK). China A-shares =
  the 6-digit code (Kweichow Moutai = 600519). Futures = Yahoo suffix (ES=F, NQ=F, GC=F,
  CL=F). Crypto = BTC-USD, ETH-USD.
  Resolve an instrument NAME to a ticker before calling: S&P 500 -> SPY or ES=F,
  Nasdaq -> QQQ or NQ=F, gold -> GLD or GC=F, crude -> CL=F.
  "GOLD" is Barrick Gold, the mining company — not the metal.

HOW TO ANSWER
- Report the whole probability distribution, not just the top class, and state the top-2
  coverage (the two highest probabilities added together).
- Give the baseline whenever you quote accuracy: 66% top-1 and 80% top-2 against a 37%
  majority-class baseline, on held-out years. At 90 minutes in: 53.5% and 68.6%.
- The day-type model is trained on S&P 500 E-mini futures. Reads on single stocks, gold and
  crypto are descriptive, not validated. Say so in the same sentence as the number.
- For Weis events, quote the measured win rate and sample size from the reference block,
  next to the 49.3% baseline. no_demand_short (45.6%) and sot_short (47.6%) measured
  REVERSED versus what the tradition teaches — never report them as bearish.
- 24h futures default to the US day session. If a response carries "session_hint", ask once
  whether the user wants the US day session or the Asian session, recommend US, remember it.

RULES
- Never present any of this as a buy or sell recommendation, and never give personalized
  investment advice. It is descriptive analytics about chart shape and volume.
- Always carry the disclaimer field through to the user.
- Never claim precision the numbers do not have. Real surviving edge here is 52-56%.
- 401 = check the key. 402 = subscription lapsed or capped. 403 = valid key, not subscribed
  to that product. 404 = symbol did not resolve; the message names a better ticker.
  Do not retry 401, 402 or 403.

What the YAML header looks like

Two fields, that is all. The description is the only part the agent reads until your question matches it, which is why it is written in the third person and says exactly when to trigger.

---
name: quantdata-daily-bias
description: Reads day-type probabilities for a stock, ETF or futures
  contract's current session from the Quant Data Brooks Daily Bias API,
  finds the most similar historical trading days, and quotes fact-checked
  price-action statistics. Use when the user asks what kind of day something
  is having, whether today looks like a trend day or a range day, for the
  daily bias or intraday character of a symbol, for historical analogs of
  today's chart shape, or whether a price-action claim such as the "80% rule"
  actually holds up in data.
---

What every one of these Skills enforces

These rules are written into each file, and they are not decoration. They are the reason a Skill is worth more than a link to an API reference.

Never present output as a buy or sell recommendation. The API returns a probability distribution and historical frequencies. It does not know your position, your account or your risk. Turning "TrendUp 0.34" into "go long" is the agent inventing advice.
Always carry the disclaimer. Every response has a disclaimer field, and it travels with the numbers.
Never claim precision the numbers do not have. Honest win rates here are 52–56%. An agent that rounds that up to "strong signal" has broken the product.
Name the baseline. 66% accuracy means nothing until you also say the majority-class baseline is 37%.
Flag out-of-distribution use. The day-type model was trained on S&P 500 E-mini futures. A read on a single stock, on gold or on crypto is descriptive, not validated — and for BTC, ETH and gold we measured that the volume-event edge is absent or reversed.

Get a key

There is no signup form and no checkout page. Email one sentence about what you are building or studying, and you get a payment link back and a key issued by hand — usually within a day.

Email quantdata@agentmail.to See pricing

Not sure what any of this means yet? Start with the learning path — it teaches the concepts first and the products second.