agentleFS
staskhstaskh/trading_skills/.claude/skills/whale-hunting/SKILL.md

whale-hunting

Skill411 starsChanged 9 days agoMITView on GitHub

Detect institutional whale activity in options for a given underlying. Use when the user asks about unusual options activity, large block trades, whale trades, or institutional options flow for a specific symbol.

  • Reads credentials

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    After running the script, present the results as follows.

    line 67
  2. Very specific:

    All timestamps and time-based calculations must use the America/New_York timezone.

    line 97

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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

Whale Hunting

Scans option chains for a given underlying to identify institutional-sized trades using a two-step approach:

  1. Crude scan (Yahoo Finance) — finds contracts with anomalous daily investment vs the rest of the chain.
  2. Precise drill-down (Massive API) — fetches per-second bars for each candidate and flags seconds with outlier dollar invested.

Instructions

Note: If uv is not installed or pyproject.toml is not found, replace uv run python with python in all commands below.

uv run python .claude/skills/whale-hunting/scripts/whale_hunting.py SYMBOL [--months N] [--date YYYY-MM-DD] [--sigma F] [--sigma-z F] [--summary]

Arguments

  • SYMBOL — Underlying ticker (e.g. AAPL, NVDA, SPY)
  • --months — Max months until option expiration to consider (default: 2)
  • --date — Trading date to analyze in YYYY-MM-DD format (default: latest trading day)
  • --sigma — Std-deviation multiplier for crude outlier threshold (default: 3.0)
  • --sigma-z — Modified Z-Score threshold for per-second small-sample detection (default: 3.5)
  • --summary — Also compute per-ticker summary and include it in the JSON output

Output

Returns JSON with:

  • underlying — The scanned symbol
  • trading_date — Date analyzed
  • source — "massive" (per-second data) or "yahoo only" (daily chain data)
  • total_whales — Total whale events found
  • total_call_invested — Sum of invested dollars in call whale events
  • total_put_invested — Sum of invested dollars in put whale events
  • call_put_ratio — Call invested / put invested (null if no puts)
  • directional_call_invested — Call invested where break-even is ≥ 1.5% above spot
  • directional_put_invested — Put invested where break-even is ≥ 1.5% below spot
  • directional_call_put_ratio — Directional call / directional put invested (null if no directional puts)
  • whales — List of whale events:
    • timestamp, ticker, type, strike, spot (underlying price at trade time), expiry
    • close, volume, transactions, invested, break_even, pct_from_spot (break-even distance from spot, %)
  • summary (present only when --summary is passed) — List of per-ticker aggregates:
    • ticker, type, strike, expiry, whale_count, total_invested, break_even

Examples

# Hunt whales for AAPL (latest trading day)
uv run python .claude/skills/whale-hunting/scripts/whale_hunting.py AAPL

# Hunt whales for NVDA on a specific date
uv run python .claude/skills/whale-hunting/scripts/whale_hunting.py NVDA --date 2026-03-13

# With per-ticker summary
uv run python .claude/skills/whale-hunting/scripts/whale_hunting.py HOOD --months 3 --summary

# Looser detection threshold
uv run python .claude/skills/whale-hunting/scripts/whale_hunting.py SPY --sigma 2.0

Reporting

After running the script, present the results as follows.

Header line:

Whale activity for {underlying} on {trading_date} — source: {source} Call flow: ${total_call_invested:,.0f} | Put flow: ${total_put_invested:,.0f} | C/P ratio: {call_put_ratio:.2f} Directional (≥1.5% B/E from spot) — Calls: ${directional_call_invested:,.0f} | Puts: ${directional_put_invested:,.0f} | C/P ratio: {directional_call_put_ratio:.2f}

When --summary was requested, render the summary array as a table:

Time (ET)TickerTypeStrikeExpiry# EventsTotal InvestedBreak Even
{timestamp}{ticker}{type}{strike}{expiry}{whale_count}${total_invested:,.0f}{break_even}

Sort by total_invested descending. For multi-event rows use the time range of first–last event (e.g. 11:46–12:33).

Interpretation guidance:

  • source: "massive" — High-confidence; per-second block trade data from Massive API
  • source: "yahoo only" — Fallback; daily-level data (Massive API key missing or no intraday data)
  • Judge direction by directional_call_put_ratio, not call_put_ratio — trades with break-even within 1.5% of spot (deep ITM stock replacement, hedges) carry no directional signal
  • Low directional C/P ratio (< 0.5) — Bearish institutional positioning
  • High directional C/P ratio (> 2.0) — Bullish institutional positioning
  • transactions: 1 — Single block trade; strongest whale signal

Requirements

  • MASSIVE_API_KEY environment variable for per-second data. Without it, falls back to Yahoo Finance daily data.

Timezone

All timestamps and time-based calculations must use the America/New_York timezone. All JSON output must include generated_at (NY time string) and data_delay fields.

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.