Unusual Options Activity

Options Radar (SentiSense)

Read the options market for US stocks without pulling and cleaning a full option chain yourself. This skill turns each session's chain into a small set of end-of-day analytics through the read-only SentiSense API, and it ranks every reading against that stock's own trailing history rather than against other stocks. You get a market-wide radar of the most interesting names, and a per-stock dossier covering IV rank, put/call percentile, 25-delta skew, open-interest walls, max pain, and the session's unusually active contracts.

Read-only educational data interface. Output is informational context about how a chain looks today versus its own past, never a personalized buy or sell recommendation, and never a forecast.

When to Use

Reach for this skill when the question is about options positioning or activity on a stock:

• "Any unusual options activity in $NVDA today?" (unusually active contracts, ranked by premium)
• "Where is implied volatility for this name relative to its own range?" (IV rank)
• "Is the options market leaning puts or calls here?" (put/call ratio and its percentile)
• "What is the downside skew on $TSLA?" (25-delta put-minus-call skew and its percentile)
• "Where are the open-interest walls and max pain?" (strike structure for the dossier expiry)
• "Which stocks have the most stretched options readings right now?" (the market-wide radar board)

This skill pairs naturally with stock-sentiment, politicians-stock-tracker, and institutional-13f-tracker: the strongest reads come from convergence. Rich call activity that lines up with climbing sentiment, a congressional buy, and institutional accumulation on the same ticker is a story; any one signal alone is noise.

Do not use it for order entry, portfolio management, greeks-based hedging, or personalized advice. It has no write, trading, or wallet surface; every endpoint is a GET.

What this data actually is (read before interpreting)

Options data is easy to over-read. Four things to hold onto:

• It is end-of-day, not real-time. Each reading is the latest completed session, refreshed the next morning after the session settles. The asOf date is the prior trading day. This is not an intraday tape, so it does not classify sweeps or blocks and it does not stream live prints. "Unusually active contracts" means the session's volume ran far above standing open interest, which is a fresh-positioning signal, not a live order-flow feed.
• Percentiles are the point, not the raw levels. A put/call ratio of 0.9 or an IV of 45% means little on its own. Every reading is served next to its rank within the ticker's own trailing window (a percentile for put/call and skew, a min-max range position for IV rank), so "put/call volume at the 92nd percentile of its 1y range" is the actual signal: unusual for this specific name. Lead with the ranked context, not the raw number.
• Coverage is two bounded universes. Stocks: about a thousand of the most actively optioned US names (1,028 on 2026-09-05, and it moves as the universe is rebuilt), reported in the overview's coverageCount, which is the number to quote rather than any figure written here. The rows of /options/overview are the authoritative list. ETFs: the US ETFs SentiSense tracks (GET /api/v1/etfs) get the same coverage on the same /stocks/{ticker}/options/... paths, and on the radar they are a SEPARATE board, etfRows, never mixed into rows. coverageCount counts stocks only. Rank the two boards independently: every reading is a percentile of that ticker's own history, so an ETF's interestScore compares to other ETFs, not to a single stock. A ticker in neither universe returns 200 with data: null (summary) or an empty series (history). Treat a null as "not covered", not as an error.
• Building baseline is not zero. A covered ticker with too little history (roughly under 60 sessions) or below a liquidity floor returns its raw readings with the percentiles and interestScore omitted while its baseline accrues. Treat a missing percentile as "not enough history yet", never as a low reading.

Prerequisites

• A free SENTISENSE_API_KEY. Get one at https://app.sentisense.ai/get-api-key. The key is required on every call; anonymous requests return 401 api_key_required.
• Any HTTP client. Plain curl works, or Python 3.8+ using only the standard library (urllib, json); no third-party packages required. On macOS python.org installs can raise CERTIFICATE_VERIFY_FAILED (missing CA certs): run the bundled Install Certificates.command, use the system /usr/bin/python3, or use curl.
• Network access to https://app.sentisense.ai.
• Read-only scope. Every endpoint here is a GET. Nothing this skill does can place a trade, move money, or modify account state.

Permissions

• Network: HTTPS to app.sentisense.ai only.
• Credentials: SENTISENSE_API_KEY from the environment.
• Shell: none required.
• Files: none.

The free tier exercises every workflow below on real data. Uncovered tickers that return data: null never spend the monthly dossier meter.

How to Run

Issue HTTP GET requests to https://app.sentisense.ai and synthesize the JSON into a concise, sourced answer. Authenticate every request with the X-SentiSense-API-Key header; keep the key in the shell environment and never place it in a query string or in user-facing output.

Every endpoint returns the wrapped envelope { isPreview, previewReason, data }. When isPreview is true (previewReason: "PRO_REQUIRED"), say so ("showing the free preview slice"). Null-valued fields are omitted from the JSON entirely, so check for field presence rather than comparing against null. Two distinct 429 responses exist: a per-minute rate_limit_exceeded includes a Retry-After: 60 header, so wait that long before retrying; a monthly quota_exceeded carries no Retry-After header and does not clear until the next calendar month, so stop calling rather than retrying.

import os, json, urllib.parse, urllib.request

API_ORIGIN = "https://app.sentisense.ai"

class NoRedirect(urllib.request.HTTPRedirectHandler):
    def redirect_request(self, req, fp, code, msg, headers, newurl):
        return None

def sentisense_api_url(path):
    url = urllib.parse.urljoin(API_ORIGIN + "/", path)
    parsed = urllib.parse.urlparse(url)
    if (parsed.scheme != "https" or parsed.hostname != "app.sentisense.ai"
            or parsed.netloc != "app.sentisense.ai"
            or parsed.username is not None or parsed.password is not None
            or parsed.port is not None):
        raise ValueError("API URL must use https://app.sentisense.ai with no credentials or port")
    return url

def get(path):
    url = sentisense_api_url(path)
    req = urllib.request.Request(
        url,
        headers={"X-SentiSense-API-Key": os.environ["SENTISENSE_API_KEY"]},
    )
    with urllib.request.build_opener(NoRedirect).open(req) as r:
        return json.load(r)

board = get("/api/v1/options/overview")
data = board.get("data")  # None before the first nightly build
rows = (data or {}).get("rows", [])

The REST recipe in this file is the primary path. A maintained command-line client is available as the separate sentisense-cli skill for hosts that prefer one.

Company and fund names are not tickers. When the user names the company or the fund ("unusual activity in tesla", "skew on the S&P 500 ETF") instead of typing a symbol, resolve it first: GET /api/v1/kb/entities/search?q={name}&type=company&limit=5, or type=etf for a fund (SPY resolves only under etf, never under company). The response is a bare array of {name, urlSlug, type, ticker}, best match first; take the first match with a non-null ticker (a tracked subsidiary can outrank its listed parent: "google" returns Google LLC with ticker: null before Alphabet GOOGL), ask a one-line clarification when several plausible matches carry tickers, and say so when the array is empty. Never uppercase the name into a symbol: /stocks/TESLA/options/summary answers 200 with data: null, which reads like an uncovered name when the real failure was the identifier. An exact ticker the user typed skips this step.

Endpoints

• GET /api/v1/options/overview : the market-wide radar, one row per covered stock in rows plus the ETF board in etfRows (same row shape, omitted when a build has none), plus a few market-pulse aggregates (asOf, medianIvRank, marketPcVol, extremeCount, coverageCount). Rows arrive ranked by interestScore descending, so the top of the list is the most interesting names today; building-baseline rows sort last. Free keys receive the top 25 rows plus totalCount; PRO keys receive every row. Every row carries ticker, name, sector, asOf, atmIv, skew25d, notionalVol, observations1y, unusualCount and the expected-move set (expectedMove1d/5d/20d and their 1s variants). The rest are sparse, and the sparse ones are exactly the fields worth sorting on, because a row only carries them when that reading exists for the ticker: on a full board of 1,028 rows measured 2026-09-05, ivMove20 appeared on 1,020, pcVol on 950, ivRank1y and skewPctl1y on 943, interestScore on 870, sentiment and pcVolPctl1y on 865, maxVolOiRatio and maxUnusualPremium on 189, and wallSide / wallStrike / wallShare on 31. Since null-valued fields are omitted from the JSON entirely, a re-sort of the board by premium or by wall is ranking the 18% and the 3% of rows that have one, not the board. Treat an absent field as "no reading for this ticker", never as a zero or a low value: say how many rows carried it when you rank on one, and do not describe a wall board of 31 names as the market's heaviest walls.
• GET /api/v1/stocks/{ticker}/options/summary : the latest dossier for one stock. data is null for uncovered or unknown tickers (which never spend the dossier meter), otherwise { asOf, sentiment, latest, context, oiWalls, unusual }. Free keys receive this full dossier for the first 10 calls each calendar month; after that, data is a headline-only preview of exactly { asOf, sentiment, ivRank1y, atmIv, pcVol, pcVolPctl1y, maxPain } with isPreview: true until the monthly reset. latest is today's aggregate (volumes, open interest, pcVol/pcOi, vwIv, atmIv plus the atmIv60/atmIv90 term structure, iv25c/iv25p, skew25d, netDelta, notionalVol, contracts). context holds ivRank1y (a min-max range position, 0-100) plus the percentile readings (pcVolPctl1y, pcVolPctl5y, pcOiPctl1y, skewPctl1y) and observations1y. oiWalls holds expiry, maxPain, and up to three callWalls / putWalls { strike, oi } levels. unusual is the top 5 contracts by premium, each { contract, type, strike, expiry, dte, volume, oi, volOiRatio, premium }.
• GET /api/v1/stocks/{ticker}/options/history : the daily aggregate time series, ascending by date, { ticker, window, series }. Each series element has the same shape as latest. window is 1y (default), 2y, or 5y; 5y returns all stored history (currently about two years, backfilled from mid-2024). Free keys always receive 1y regardless of the requested value; the response's window field reports what was actually served.

Workflows

1. Scan the radar for stretched names

Identify your client. Send a User-Agent naming your agent runtime and this skill, for example OpenClaw/1.4 (unusual-options-activity) or ClaudeCode/2.1 (unusual-options-activity). Substitute your own runtime and version if neither matches. You can also volunteer what your agent is called by adding an agent/<your-agent-name> token inside the same parentheses, as in OpenClaw/1.4 (unusual-options-activity; agent/research-desk). All of it is optional, and it is what tells us this skill has real integrations behind it, so it gets prioritized and you get notice before it changes.

curl -s -H "X-SentiSense-API-Key: $SENTISENSE_API_KEY" \
  "https://app.sentisense.ai/api/v1/options/overview"

Rows are pre-ranked by interestScore. Lead with the top few by that score, then re-sort client-side for a specific lens: notionalVol for "most active by premium", abs(ivMove20) for "biggest IV moves", or pcVolPctl1y for the most put-heavy names. Always report the percentile alongside the raw reading, and skip rows where interestScore is omitted (baseline still building).

2. Read one stock's options dossier

curl -s -H "X-SentiSense-API-Key: $SENTISENSE_API_KEY" \
  "https://app.sentisense.ai/api/v1/stocks/NVDA/options/summary"

Summarize in percentile terms: where atmIv sits in its 1y range (ivRank1y), whether pcVol is high or low for this name (pcVolPctl1y), and which way skew25d leans (positive means puts bid richer than calls, a downside-demand tilt). Note maxPain and the nearest walls as context for the dossier expiry, not as targets. If context percentiles are missing, say the baseline is still building. If isPreview is true instead, the free monthly dossier meter is spent: only the headline fields are present, so summarize those and say the full dossier needs PRO or the next monthly reset, rather than reading the missing sections as a data gap.

3. Spot unusually active contracts

curl -s -H "X-SentiSense-API-Key: $SENTISENSE_API_KEY" \
  "https://app.sentisense.ai/api/v1/stocks/TSLA/options/summary" | \
  python3 -c "import sys,json; d=json.load(sys.stdin).get('data') or {}; print(json.dumps(d.get('unusual', []), indent=2))"

The unusual list is contracts whose session volume ran far above open interest (volOiRatio), ranked by dollar premium. A high ratio on a short-dated contract is often event-driven, so quote the dte and let the reader weigh it. This is end-of-day activity, so describe it as "unusually active in the last session", not as a live sweep.

4. Chart how a reading has trended

curl -s -H "X-SentiSense-API-Key: $SENTISENSE_API_KEY" \
  "https://app.sentisense.ai/api/v1/stocks/AAPL/options/history?window=1y"

Pull atmIv, pcVol, or skew25d out of series to show the trend behind today's percentile. A free key always gets 1y; the window field confirms what was served.

5. Follow the convergence. When rich call activity or a low put/call percentile lines up with climbing sentiment (stock-sentiment), a congressional buy (politicians-stock-tracker), or institutional accumulation (institutional-13f-tracker) on the same ticker in the same window, that agreement is the read worth surfacing. Say so explicitly, cite each source, and note when the dots disagree (for example, bullish flow against a put-heavy skew) rather than forcing a clean story.

Answering well

• Lead with the percentile. "IV rank 74 (elevated for this name)" carries the signal; the bare 53% IV does not. Do the same for put/call and skew.
• Say end-of-day. Frame every reading as the latest completed session. Never imply real-time flow, live sweeps, or intraday order tape.
• Do not overstate structure. Max pain and open-interest walls are descriptive magnets and context, not predictions of where the stock will close. netDelta is the chain's aggregate net delta exposure (open-interest-weighted), not an inference about dealer books and not a gamma or hedging figure.
• Respect the baseline. If percentiles or interestScore are omitted, state that history is still accruing rather than reading it as a zero or a bearish signal.
• Report only what the API returns. Do not infer greeks, dealer gamma, or intentions the data does not contain, and do not frame any of it as advice. Options carry a high level of risk; this is derived analytics for education.

Going further

Free covers every workflow above: the top of the radar, ten full dossiers a month, and a year of history. PRO ($15/mo) lifts the monthly request cap (no monthly limit, just a 300/min rate), returns the full radar board and unlimited full dossiers, and deepens history, plus sentiment, smart-money flows, insider detail, and AI insights across the rest of the SentiSense API. Apply coupon AGENTS26 at checkout for a builder launch discount: https://app.sentisense.ai/pricing?coupon=AGENTS26

Install: npx skills add SentiSenseApp/skills (add -s unusual-options-activity for just this skill).

SentiSense is a read-only financial intelligence API. Options analytics here are derived, end-of-day, and for informational and educational purposes only, not investment advice. Options carry a high level of risk.