starintel-labs/starintel-gpt-auto-dig

Fork 0

[GH-#99] Create `oil-futures`: 2026 oil-market corruption events and deep person histories #997

New issue

Open

opened 2026-09-08 17:05:44 +00:00 by nsaspy · 0 comments

nsaspy commented

2026-09-08 17:05:44 +00:00

Owner

Mirrored from GitHub https://github.com/lost-rob0t/starintel-gpt-auto-dig/issues/99 (GitHub is authoritative for this item).

Objective

Create a validated StarIntel v0.9.0 dataset named oil-futures investigating corruption, manipulation, misuse of nonpublic information, sanctions evasion, benchmark abuse, procurement fraud, and investor fraud affecting oil and fuel markets.

The event ledger is strictly limited to events occurring in calendar year 2026. Historical material may be collected only to build deep profiles of people and organizations implicated, scrutinized, sanctioned, charged, convicted, or materially connected to a 2026 event.

This is not a generic oil-industry dataset and not merely a futures dataset. Build an evidence graph tracing:

nonpublic decision or market-moving event → information-access chain → timed transaction cluster → exchange/account/firm → responsible or controlling people → regulatory response → prior professional/enforcement history

and, outside futures:

physical commodity / benchmark / cargo / sanctions / procurement / offering → legal entity → responsible people → transaction network → enforcement or adjudicated outcome

Dataset

oil-futures

Initial run path:

digs/oil-futures/2026-07-31-2026-market-corruption-root/

Hard temporal rule

A record belongs in the 2026 corruption-event ledger only when the suspicious trade, enforcement action, charge, conviction, sentence, sanctions designation, settlement, forfeiture filing, official investigation, or other material event occurred from 2026-01-01 through 2026-12-31.

Historical facts are allowed only in:

person career and affiliation timelines;
organization ownership and control histories;
prior enforcement, litigation, sanctions, compliance, and disciplinary histories;
prior trading or benchmark-manipulation patterns used as sourced context;
predecessor/successor and fund lineage.

Do not mislabel a 2018 trade or 2024 enforcement order as a 2026 corruption event.

Central 2026 suspicious-trading cluster

March 23, 2026

Reuters found short bets across Brent, WTI, ICE gasoil, and U.S. gasoline totaling about $2.2 billion at 10:49–10:50 GMT, before President Donald Trump announced at 11:05 GMT that threatened attacks on Iranian power infrastructure would be delayed. Reuters could not identify who placed the bets. Crude and fuel prices fell sharply after the announcement.

Source:

Reuters, 2026-05-07: https://www.reuters.com/sustainability/boards-policy-regulation/oil-price-bets-ahead-iran-war-news-totalled-7-billion-reporting-shows-2026-05-07/

The Wall Street Journal reported that the CFTC was examining whether someone with advance knowledge traded or leaked the information and identified firms whose activity appeared in records:

Qube Research & Technologies — about $5 million adjusted gain;
Forza Fund Ltd. — about $10 million adjusted gain;
TOTSA TotalEnergies Trading SA — about $200,000 adjusted gain.

The firms have not been accused of wrongdoing. Some firms reportedly attributed trades to a Semafor headline published before Trump’s post. Preserve that alternative explanation and obtain the exact publication timestamp and market-data arrival path.

Additional transaction-context firms reported in records include:

Jane Street — approximately $19 million adjusted gain;
Jump Trading — approximately $15 million adjusted loss;
Virtu Financial — approximately $3 million adjusted gain;
Shell trading desks — seven-figure adjusted gain;
IMC Chicago — seven-figure adjusted gain;
Paragon Trading Partners — approximately $3 million adjusted gain;
TTG Capital — approximately $1 million adjusted gain;
Tower Research Capital — more than $3 million adjusted gain.

These context firms are not automatically suspects. Distinguish market makers, systematic strategies, manual directional trades, spread trades, hedges, and related-asset positions.

Source:

WSJ, 2026-05-20: https://www.wsj.com/finance/regulation/flurry-of-suspicious-oil-trades-worth-800-million-triggers-regulatory-probe-71e959ce

April 7, 2026

Reuters found about $2.12 billion in oil and gasoline sell orders at 19:44–19:45 GMT during thin post-settlement trading, shortly before Trump announced a two-week ceasefire with Iran. Preserve contract month, exchange, product, lot count, price, aggressor side, and announcement timestamp.

April 17, 2026

Reuters found nearly $2 billion in Brent, WTI, gasoil, and gasoline futures sold at 12:24–12:25 GMT, before Iranian Foreign Minister Abbas Araghchi announced that the Strait of Hormuz would reopen and before related U.S. and Iranian posts.

April 21, 2026

Reuters found about $830 million in Brent and WTI contracts sold roughly 15 minutes before Trump extended the ceasefire.

Combined scope

Reuters calculated as much as $7 billion in well-timed short positions across the four dates after expanding from front-month crude into longer-dated crude, diesel, and gasoline contracts. Reuters said it could not establish who placed the bets or whether the accounts originated in the United States or elsewhere.

The CFTC and CME were reported to be examining the trades; Reuters reported that the CFTC had not publicly confirmed its investigation at that time. The Justice Department was also reported to be investigating. Treat investigation status as time-sensitive and refresh from official sources before publishing.

Additional 2026 suspicious oil-trading event

May 6, 2026

Investigate the pre-dawn crude-futures volume surge before reporting about progress in U.S.–Iran talks. MarketWatch reported roughly 17,300 WTI contracts, approximately $1.7 billion notional, traded from 3:50–4:10 a.m. ET before an Axios report around 4:50 a.m. ET. Obtain CME tick data, contract months, trade direction, price impact, market-depth changes, account identifiers where legally/publicly available, and exact news timestamps.

Source:

MarketWatch: https://www.marketwatch.com/story/traders-point-to-suspicious-activity-in-the-oil-market-on-wednesday-3a6821d9

Congressional and executive-branch response

Warren–Whitehouse request

On April 10, 2026, Senators Elizabeth Warren and Sheldon Whitehouse publicly requested a CFTC investigation into March 23 and April 7 trading and asked whether material nonpublic government information had been misappropriated.

Source:

Senate Banking Committee: https://www.banking.senate.gov/newsroom/minority/warren-whitehouse-probe-suspicious-oil-trades-surrounding-trump-iran-announcements

Collect:

the complete signed letter and attachments;
any CFTC response due April 30;
hearing testimony and follow-up correspondence;
congressional requests to DOJ, SEC, CME, ICE, White House Counsel, OGE, and inspectors general;
any FOIA releases, subpoenas, referrals, enforcement filings, or closed-without-action statements.

White House staff warning

Reuters and WSJ reported that White House staff were warned against using nonpublic information for prediction-market or futures bets after attention focused on the March 23 trades. Obtain the exact memorandum/email, sender, recipients or distribution scope, issue date, legal citations, ethics-office involvement, and any preservation notice. The warning is evidence of an internal compliance response, not evidence that a specific employee traded.

Priority organizations and people for deep historical profiles

Qube Research & Technologies

Create the exact UK legal entity, funds, managers, affiliates, former names, and regulatory registrations.

Verified initial structure:

Qube Research & Technologies Limited, UK company 09867306;
incorporated in 2015 under the former name Credit Suisse Quantitative and Systematic Asset Management Limited;
Credit Suisse Asset Management (UK) Holding Limited held at least 75% of shares/voting rights and director-appointment control until January 1, 2018;
current Companies House filing says there is no registrable person or relevant legal entity with significant control;
SEC filings identify Pierre-Yves Morlat as CEO and QRT as investment manager to QRT funds.

Primary sources:

Companies House overview: https://find-and-update.company-information.service.gov.uk/company/09867306
Officers: https://find-and-update.company-information.service.gov.uk/company/09867306/officers
PSC: https://find-and-update.company-information.service.gov.uk/company/09867306/persons-with-significant-control
QRT official site: https://www.qube-rt.com/
SEC adviser record: https://adviserinfo.sec.gov/firm/summary/304552

Initial people:

Pierre-Yves Morlat — CEO/director;
Laurent Laizet — director/founding-era officer;
Stuart David Brown — COO/director and public spokesman on the 2026 trading report;
Christina Wilgress — director and signatory/attorney-in-fact on SEC filings;
Marcus Knight — director;
Stephen Leslie Foster — director;
Claire Michelle Ritchie — appointed director April 23, 2026;
former Credit Suisse officers and owners tied to the 2015–2018 transition.

For each person, recover:

complete employment and board timeline;
exact Credit Suisse/QSAM role and dates;
fund-management, trading, risk, compliance, and technology responsibility;
legal entities and funds under authority;
regulatory registrations and disclosures;
prior enforcement, litigation, disciplinary records, and public controversies;
political donations, government advisory roles, lobbying, and material board interlocks only when sourced;
whether the person had any documented connection to the March 23 trading desk or strategy. Do not infer responsibility from title alone.

Forza Fund Ltd. and Metabit Trading

Resolve Forza before asserting jurisdiction, asset class, ownership, directors, or investment strategy. Current public reporting links Forza to Metabit Trading, a China-based quantitative trading firm, but public beneficial ownership remains unclear.

Initial Metabit people to verify from authoritative Chinese registry, AMAC/private-fund records, official corporate material, and archived websites:

Pengfei Gao;
Ran Xian;
Jiang Renfeng;
any Forza directors, investment managers, traders, compliance personnel, fund administrators, custodians, prime brokers, or beneficial owners.

Required work:

exact Forza legal name, jurisdiction, registration number, registered office, incorporation date, directors, shareholders, beneficial owners, fund administrator, auditor, prime broker, investment manager, and regulatory status;
precise legal and operational relation between Forza and Metabit;
Chinese private-fund manager registration, products, scale history, affiliates, offshore vehicles, and international futures access;
people with authority over oil-futures strategies in March–May 2026;
historical enforcement, exchange discipline, litigation, sanctions, or compliance issues;
do not copy the unsupported description of Forza as a “real-estate financing firm” without registry evidence.

TOTSA TotalEnergies Trading SA

Resolve the Swiss legal entity and current/historical governance.

Initial entity and people:

TOTSA TotalEnergies Trading SA, Geneva UID CHE-103.194.266;
Rahim Azouni — current president/CEO in Swiss registry data;
Thomas Waymel — prior president and long-serving TotalEnergies Trading & Shipping leader;
Etienne Dinner;
Franck Trochet;
Frédéric Agnès;
Alev Gülaçar;
Patrick Pouyanné — TotalEnergies SE chairman/CEO, as parent-company governance context, not presumed desk responsibility.

Primary sources:

legal notice: https://trading.totalenergies.com/en/legal-notices/
CFTC 2024 enforcement: https://www.cftc.gov/PressRoom/PressReleases/8953-24
TotalEnergies annual reports: https://totalenergies.com/investors/publications-and-regulated-information/regulated-information/annual-financial-reports

Historical priority:

In 2024, the CFTC ordered TOTSA to pay $48 million for attempted manipulation of EBOB-linked futures. The order found that in March 2018 TOTSA sold large volumes of physical EBOB at below indicated bids while holding a large short EBOB-linked futures position, and that potentially relevant WhatsApp evidence was unavailable because communications were not timely preserved or produced. Preserve Commissioner Caroline Pham’s dissent and TOTSA’s defense/counterevidence.

This prior adjudicated settlement is material background, but it does not establish that TOTSA’s 2026 crude trades were improper.

2026 non-futures and physical-market corruption/enforcement branch

Iranian oil shipping, shell companies, and sanctions evasion

Create a dedicated network around Mohammad Hossein Shamkhani and recurse into each person named by DOJ/Treasury in 2026 actions.

2026 root events:

March 6 DOJ civil-forfeiture complaints against more than $15.3 million allegedly connected to an Iranian oil distribution and shipping network;
April 15 Treasury sanctions against individuals, entities, and vessels in the Shamkhani network;
July 14 Treasury action targeting more than 50 additional individuals, entities, and vessels.

Primary sources:

DOJ: https://www.justice.gov/opa/pr/united-states-files-civil-forfeiture-complaints-against-15m-funds-allegedly-linked-iranian
Treasury, April 15: https://home.treasury.gov/news/press-releases/sb0443
Treasury, July 14: https://home.treasury.gov/news/press-releases/sb0562

Initial people include:

Mohammad Hossein Shamkhani;
Hossein Ghorbani Zahed;
Mohammad Reza Rahbar Madani;
Ali Rakhbarmadani;
Martin Austin Kaalund;
Alessandra Ronco;
Asghar Aghili Dehkordi;
Jijin George;
Behzad Moghadas;
Gautam Vishavdeep;
Maksim Tsernosjov;
every additional individual named in the 2026 complaints/designations.

For each, recover corporate directorships, beneficial ownership, passports/nationalities only as officially published, shipping and commodities roles, vessel/company edges, sanctions history, litigation, prior employers, Russian-oil price-cap activity, and chronological changes. Treat DOJ complaint allegations as allegations and OFAC designations as official administrative actions, not criminal convictions.

Ali Ansari and shadow exchange houses

Create the July 10, 2026 Treasury event and recursively map Ali Ansari, named exchange houses, banks, shell companies, overseas assets, and beneficiaries. Treasury alleged that Ansari institutionalized large-scale embezzlement and used a global asset network to benefit Iranian regime elites.

Source:

https://home.treasury.gov/news/press-releases/sb0558

Adani Enterprises Iranian LPG settlement

Create the May 18, 2026 OFAC settlement event:

Adani Enterprises Limited agreed to pay $275 million to resolve potential civil liability for 32 apparent Iran-sanctions violations;
OFAC said the conduct was egregious and not voluntarily self-disclosed;
transactions involved LPG represented as Omani or Iraqi but allegedly originating in Iran and about $192.1 million in U.S.-dollar payments.

Source:

https://ofac.treasury.gov/recent-actions/20260518

Recurse into the exact business unit, procurement/trading decision chain, compliance officers, Dubai supplier, shipping companies, vessels, banks, counterparties, investigators, and remedial actions. Do not assign personal culpability to Gautam Adani or any executive without evidence in the settlement or related records.

Fuel procurement fraud

Create 2026 conviction and sentencing events for Jasen Butler, owner of Independent Marine Oil Services LLC:

January 16 jury conviction on 34 felony counts;
April 8 sentence of 60 months;
more than $4.5 million in falsified fuel-related expenses submitted through military vessel procurement.

Sources:

Deep-profile Butler, his company, employees, counterparties, SEA Card contracting chain, vessels, ports, invoices, shell entities, assets, prior businesses, and related litigation.

Biofuel-credit fraud

Create the May 29, 2026 sentencing event for Christopher Burdett, including General Manager Royce Gillham, the company, fraudulent Renewable Identification Numbers/credits, overstated production, tax-credit claims, restitution, and prior case history.

Source:

https://www.justice.gov/opa/pr/florida-man-sentenced-18-months-prison-biofuel-fraud-conspiracy

Oil-and-gas securities and offering cases

Create separate 2026 enforcement-event records, preserving that many underlying schemes began before 2026:

Kevin N. Richards — April 7 final judgment involving alleged unregistered oil-and-gas offerings and undisclosed conflicts;
David P. Ortiz / DaveGlo Investment Group — April 27 final judgments;
Michael Bowen / Cannon Operating — June 8 final consent judgment;
additional SEC oil-and-gas offering cases filed, adjudicated, or settled in 2026.

Sources:

Keep investor-offering fraud separate from commodity-price manipulation.

Required recursive execution

Run a root pass plus six explicit recursive passes. Every pass must create a research-pass record documenting inputs, queries, sources, records created, selected/deferred targets, confidence, unresolved identities, contradictions, and next frontier.

Pass 0 — 2026 event ledger and market timeline

create one event per suspicious trading window, investigation, enforcement action, sanction, conviction, sentence, settlement, or forfeiture filing;
normalize every timestamp to UTC while preserving original timezone;
record announcement creation time, publication time, wire/headline distribution time, social-post time, exchange timestamp, market reaction, and source retrieval time separately;
obtain contract symbols, months, lot counts, notional estimates, prices, volume baselines, open interest, market depth, aggressor side, and related-product trades;
distinguish observed trade facts from profit estimates and legal allegations.

Pass 1 — Account, firm, and execution-chain attribution

identify clearing members, FCMs, exchange members, brokers, market makers, funds, desks, strategy owners, and Tag 50/account identifiers from public records, court filings, regulatory requests, or later disclosures;
map Qube, Forza/Metabit, TOTSA, and all context firms;
distinguish automated headline response from directional discretionary trading;
compare suspicious-window activity against each firm’s ordinary baseline;
identify related positions in options, swaps, ETFs, equities, freight, crack spreads, calendar spreads, gasoline, diesel, currencies, prediction markets, and physical cargoes;
never infer the ultimate beneficial trader solely from an exchange member or clearing firm.

Pass 2 — Deep person histories

For every material person selected from Passes 0–1:

create a complete professional-role timeline;
map legal entities, funds, desks, boards, partnerships, ownership, signing authority, and compliance responsibility;
retrieve regulatory registrations, disciplinary history, litigation, sanctions, prior settlements, testimony, political/government roles, and material interlocks;
identify exact evidence linking or not linking the person to the 2026 event;
add an explicit no-direct-evidence or equivalent sourced claim when a person is included only because of governance position;
do not collect private addresses, personal phone numbers, family details, or unrelated personal information.

Pass 3 — Government-information access chain

For each March/April announcement:

reconstruct who drafted, reviewed, approved, scheduled, transmitted, translated, briefed, or technically published the decision;
use official schedules, visitor logs where public, congressional testimony, press-pool reports, archived posts, FOIA releases, ethics records, and staff biographies;
map Donald Trump and Abbas Araghchi as announcement/decision nodes;
identify U.S. and Iranian officials only where evidence places them in the relevant decision process;
distinguish access opportunity from evidence of a leak;
search for communications between government personnel and traders, donors, family businesses, lobbyists, funds, energy companies, or media only through lawful public records.

Pass 4 — Physical benchmarks, cargoes, and energy-trading history

recurse from TOTSA’s prior EBOB manipulation settlement into responsible desks, people, benchmarks, brokers, Argus methodology, and controls;
inspect 2026 Platts Dubai/Oman/Murban pricing-window concentration, cargo trading, benchmark participation, and physical/derivative alignment without labeling concentrated legitimate trading as corruption absent evidence;
map comparable adjudicated energy-market cases involving Trafigura, Vitol, Glencore, BP, Shell, TotalEnergies, and others only as historical profiles attached to 2026 entities or people;
separate Platts, Argus, exchange, OTC, physical-cargo, freight, storage, and derivatives mechanisms.

Pass 5 — Sanctions, procurement, credits, and offering fraud

recursively expand Shamkhani, Ansari, Adani LPG, Butler, Burdett, Richards, Ortiz, Bowen, and newly discovered 2026 cases;
map beneficial owners, directors, operators, traders, compliance personnel, counterparties, shell companies, vessels, banks, contracts, invoices, and proceeds;
distinguish sanctions designation, civil complaint, settlement, conviction, sentence, and allegation;
create cross-dataset links where people or organizations overlap Iran, Russia, Trump, defense procurement, TotalEnergies, shipping, commodities, or existing corruption datasets.

Pass 6 — Regulators, surveillance gaps, and accountability

profile CFTC Chairman Michael S. Selig, enforcement leadership, relevant commissioners, CME/ICE surveillance leadership, DOJ prosecutors, congressional investigators, White House Counsel/ethics officials, and inspectors general;
obtain the CFTC response to Congress and current probe status;
assess 2026 staffing cuts, reporting delays, no-action letters, large-trader reporting changes, physical commodity swaps reporting, 24/7 crude-futures proposals, and enforcement capacity as regulatory context;
do not claim a rule change caused or enabled the trades without evidence;
map conflicts, recusals, revolving-door employment, lobbying, and political contributions only from authoritative filings;
preserve non-enforcement explanations and closed investigations.

Canonical target selection

After each pass, run the repository target selector with a pass-specific query:

python3 scripts/starintel.py select-targets \
  --query "2026 oil futures corruption market manipulation" \
  --limit 20 \
  --emit-documents \
  --output recursive-targets-pass-N.jsonl

Selection priority:

person with documented control, decision access, trade authority, beneficial ownership, or enforcement relevance;
firm/account linked to a suspicious transaction cluster;
direct counterparty, broker, exchange, FCM, price-reporting agency, cargo, vessel, bank, or shell-company edge;
adjudicated prior conduct materially relevant to a 2026 entity/person;
high-value unresolved identity with a concrete primary-source acquisition path.

Do not select a person merely because they work for a large firm.

Required document classes

Use exact dtypes exposed by starintel_doc; at minimum represent:

dataset-manifest;
research-pass;
event;
org;
person;
source;
claim;
relation;
investigation target/unresolved endpoint;
financial/market observation using the closest declared schema;
vessel, contract, product/system, or asset using exact available dtypes.

Required outputs

README.md;
sources.md;
manifest.json;
canonical starintel-documents.jsonl;
Pass 0 through Pass 6 research-pass records;
recursive-targets-pass-1.jsonl through recursive-targets-pass-6.jsonl;
2026 corruption-event ledger;
millisecond/second-resolution suspicious-trade timeline where data permits;
firm/account/clearing/exchange attribution matrix;
deep person dossiers and role timelines;
physical benchmark and cargo map;
sanctions/shipping/shell-company network;
enforcement-status matrix;
source-to-claim coverage table;
contradictions and alternative-explanations table;
unresolved beneficial-owner/account queue;
explicit next-pass targets.

Evidence and accusation rules

The actual insiders, if any, are not publicly identified at dataset creation time.
A profitable or well-timed trade is not proof of insider trading.
Being named as a firm of interest is not an accusation or finding.
An algorithm reacting to a public headline is a viable alternative hypothesis and must be tested with timestamped data.
Market-maker gross or adjusted gains may not reflect the complete portfolio, hedges, spreads, or related-asset positions.
Governance title does not establish knowledge of, approval of, or participation in a trade.
Preserve government, company, commissioner, defendant, and defense explanations alongside allegations.
Label Senate statements as attributed political/oversight claims, not adjudicated facts.
Preserve exact legal posture: inquiry, investigation, request, allegation, complaint, designation, settlement, conviction, sentence, appeal, or closed matter.
Use public professional and corporate records only. No doxxing or private-person data.

Validation and Git flow

Start from main.
Create branch agent/oil-futures-2026-corruption.
Inspect executable schemas before drafting documents:

python3 scripts/starintel.py types
python3 scripts/starintel.py schema --dtype <dtype>

Do not hand-write normalized records under db/.
Write through scripts/create-db-document.py or import validated JSONL through scripts/starintel.py import.
Run:

python3 scripts/validate-for-merge.py --site

Open a draft PR only after the local gate succeeds.
Keep it draft until every required check passes on the current head.
Never weaken validation, hide unsupported claims in extensions, or merge invalid documents.

Mirrored from GitHub https://github.com/lost-rob0t/starintel-gpt-auto-dig/issues/99 (GitHub is authoritative for this item). --- ## Objective Create a validated StarIntel v0.9.0 dataset named **`oil-futures`** investigating corruption, manipulation, misuse of nonpublic information, sanctions evasion, benchmark abuse, procurement fraud, and investor fraud affecting oil and fuel markets. The event ledger is **strictly limited to events occurring in calendar year 2026**. Historical material may be collected only to build deep profiles of people and organizations implicated, scrutinized, sanctioned, charged, convicted, or materially connected to a 2026 event. This is not a generic oil-industry dataset and not merely a futures dataset. Build an evidence graph tracing: `nonpublic decision or market-moving event → information-access chain → timed transaction cluster → exchange/account/firm → responsible or controlling people → regulatory response → prior professional/enforcement history` and, outside futures: `physical commodity / benchmark / cargo / sanctions / procurement / offering → legal entity → responsible people → transaction network → enforcement or adjudicated outcome` ## Dataset `oil-futures` Initial run path: `digs/oil-futures/2026-07-31-2026-market-corruption-root/` ## Hard temporal rule A record belongs in the **2026 corruption-event ledger** only when the suspicious trade, enforcement action, charge, conviction, sentence, sanctions designation, settlement, forfeiture filing, official investigation, or other material event occurred from **2026-01-01 through 2026-12-31**. Historical facts are allowed only in: - person career and affiliation timelines; - organization ownership and control histories; - prior enforcement, litigation, sanctions, compliance, and disciplinary histories; - prior trading or benchmark-manipulation patterns used as sourced context; - predecessor/successor and fund lineage. Do not mislabel a 2018 trade or 2024 enforcement order as a 2026 corruption event. ## Central 2026 suspicious-trading cluster ### March 23, 2026 Reuters found short bets across Brent, WTI, ICE gasoil, and U.S. gasoline totaling about **$2.2 billion** at 10:49–10:50 GMT, before President Donald Trump announced at 11:05 GMT that threatened attacks on Iranian power infrastructure would be delayed. Reuters could not identify who placed the bets. Crude and fuel prices fell sharply after the announcement. Source: - Reuters, 2026-05-07: https://www.reuters.com/sustainability/boards-policy-regulation/oil-price-bets-ahead-iran-war-news-totalled-7-billion-reporting-shows-2026-05-07/ The Wall Street Journal reported that the CFTC was examining whether someone with advance knowledge traded or leaked the information and identified firms whose activity appeared in records: - **Qube Research & Technologies** — about $5 million adjusted gain; - **Forza Fund Ltd.** — about $10 million adjusted gain; - **TOTSA TotalEnergies Trading SA** — about $200,000 adjusted gain. The firms have **not been accused of wrongdoing**. Some firms reportedly attributed trades to a Semafor headline published before Trump’s post. Preserve that alternative explanation and obtain the exact publication timestamp and market-data arrival path. Additional transaction-context firms reported in records include: - Jane Street — approximately $19 million adjusted gain; - Jump Trading — approximately $15 million adjusted loss; - Virtu Financial — approximately $3 million adjusted gain; - Shell trading desks — seven-figure adjusted gain; - IMC Chicago — seven-figure adjusted gain; - Paragon Trading Partners — approximately $3 million adjusted gain; - TTG Capital — approximately $1 million adjusted gain; - Tower Research Capital — more than $3 million adjusted gain. These context firms are not automatically suspects. Distinguish market makers, systematic strategies, manual directional trades, spread trades, hedges, and related-asset positions. Source: - WSJ, 2026-05-20: https://www.wsj.com/finance/regulation/flurry-of-suspicious-oil-trades-worth-800-million-triggers-regulatory-probe-71e959ce ### April 7, 2026 Reuters found about **$2.12 billion** in oil and gasoline sell orders at 19:44–19:45 GMT during thin post-settlement trading, shortly before Trump announced a two-week ceasefire with Iran. Preserve contract month, exchange, product, lot count, price, aggressor side, and announcement timestamp. ### April 17, 2026 Reuters found nearly **$2 billion** in Brent, WTI, gasoil, and gasoline futures sold at 12:24–12:25 GMT, before Iranian Foreign Minister **Abbas Araghchi** announced that the Strait of Hormuz would reopen and before related U.S. and Iranian posts. ### April 21, 2026 Reuters found about **$830 million** in Brent and WTI contracts sold roughly 15 minutes before Trump extended the ceasefire. ### Combined scope Reuters calculated as much as **$7 billion** in well-timed short positions across the four dates after expanding from front-month crude into longer-dated crude, diesel, and gasoline contracts. Reuters said it could not establish who placed the bets or whether the accounts originated in the United States or elsewhere. The CFTC and CME were reported to be examining the trades; Reuters reported that the CFTC had not publicly confirmed its investigation at that time. The Justice Department was also reported to be investigating. Treat investigation status as time-sensitive and refresh from official sources before publishing. ## Additional 2026 suspicious oil-trading event ### May 6, 2026 Investigate the pre-dawn crude-futures volume surge before reporting about progress in U.S.–Iran talks. MarketWatch reported roughly 17,300 WTI contracts, approximately $1.7 billion notional, traded from 3:50–4:10 a.m. ET before an Axios report around 4:50 a.m. ET. Obtain CME tick data, contract months, trade direction, price impact, market-depth changes, account identifiers where legally/publicly available, and exact news timestamps. Source: - MarketWatch: https://www.marketwatch.com/story/traders-point-to-suspicious-activity-in-the-oil-market-on-wednesday-3a6821d9 ## Congressional and executive-branch response ### Warren–Whitehouse request On April 10, 2026, Senators Elizabeth Warren and Sheldon Whitehouse publicly requested a CFTC investigation into March 23 and April 7 trading and asked whether material nonpublic government information had been misappropriated. Source: - Senate Banking Committee: https://www.banking.senate.gov/newsroom/minority/warren-whitehouse-probe-suspicious-oil-trades-surrounding-trump-iran-announcements Collect: - the complete signed letter and attachments; - any CFTC response due April 30; - hearing testimony and follow-up correspondence; - congressional requests to DOJ, SEC, CME, ICE, White House Counsel, OGE, and inspectors general; - any FOIA releases, subpoenas, referrals, enforcement filings, or closed-without-action statements. ### White House staff warning Reuters and WSJ reported that White House staff were warned against using nonpublic information for prediction-market or futures bets after attention focused on the March 23 trades. Obtain the exact memorandum/email, sender, recipients or distribution scope, issue date, legal citations, ethics-office involvement, and any preservation notice. The warning is evidence of an internal compliance response, not evidence that a specific employee traded. ## Priority organizations and people for deep historical profiles ### Qube Research & Technologies Create the exact UK legal entity, funds, managers, affiliates, former names, and regulatory registrations. Verified initial structure: - Qube Research & Technologies Limited, UK company 09867306; - incorporated in 2015 under the former name **Credit Suisse Quantitative and Systematic Asset Management Limited**; - Credit Suisse Asset Management (UK) Holding Limited held at least 75% of shares/voting rights and director-appointment control until January 1, 2018; - current Companies House filing says there is no registrable person or relevant legal entity with significant control; - SEC filings identify **Pierre-Yves Morlat** as CEO and QRT as investment manager to QRT funds. Primary sources: - Companies House overview: https://find-and-update.company-information.service.gov.uk/company/09867306 - Officers: https://find-and-update.company-information.service.gov.uk/company/09867306/officers - PSC: https://find-and-update.company-information.service.gov.uk/company/09867306/persons-with-significant-control - QRT official site: https://www.qube-rt.com/ - SEC adviser record: https://adviserinfo.sec.gov/firm/summary/304552 Initial people: - Pierre-Yves Morlat — CEO/director; - Laurent Laizet — director/founding-era officer; - Stuart David Brown — COO/director and public spokesman on the 2026 trading report; - Christina Wilgress — director and signatory/attorney-in-fact on SEC filings; - Marcus Knight — director; - Stephen Leslie Foster — director; - Claire Michelle Ritchie — appointed director April 23, 2026; - former Credit Suisse officers and owners tied to the 2015–2018 transition. For each person, recover: - complete employment and board timeline; - exact Credit Suisse/QSAM role and dates; - fund-management, trading, risk, compliance, and technology responsibility; - legal entities and funds under authority; - regulatory registrations and disclosures; - prior enforcement, litigation, disciplinary records, and public controversies; - political donations, government advisory roles, lobbying, and material board interlocks only when sourced; - whether the person had any documented connection to the March 23 trading desk or strategy. Do not infer responsibility from title alone. ### Forza Fund Ltd. and Metabit Trading Resolve Forza before asserting jurisdiction, asset class, ownership, directors, or investment strategy. Current public reporting links Forza to **Metabit Trading**, a China-based quantitative trading firm, but public beneficial ownership remains unclear. Initial Metabit people to verify from authoritative Chinese registry, AMAC/private-fund records, official corporate material, and archived websites: - Pengfei Gao; - Ran Xian; - Jiang Renfeng; - any Forza directors, investment managers, traders, compliance personnel, fund administrators, custodians, prime brokers, or beneficial owners. Required work: - exact Forza legal name, jurisdiction, registration number, registered office, incorporation date, directors, shareholders, beneficial owners, fund administrator, auditor, prime broker, investment manager, and regulatory status; - precise legal and operational relation between Forza and Metabit; - Chinese private-fund manager registration, products, scale history, affiliates, offshore vehicles, and international futures access; - people with authority over oil-futures strategies in March–May 2026; - historical enforcement, exchange discipline, litigation, sanctions, or compliance issues; - do not copy the unsupported description of Forza as a “real-estate financing firm” without registry evidence. ### TOTSA TotalEnergies Trading SA Resolve the Swiss legal entity and current/historical governance. Initial entity and people: - TOTSA TotalEnergies Trading SA, Geneva UID CHE-103.194.266; - Rahim Azouni — current president/CEO in Swiss registry data; - Thomas Waymel — prior president and long-serving TotalEnergies Trading & Shipping leader; - Etienne Dinner; - Franck Trochet; - Frédéric Agnès; - Alev Gülaçar; - Patrick Pouyanné — TotalEnergies SE chairman/CEO, as parent-company governance context, not presumed desk responsibility. Primary sources: - legal notice: https://trading.totalenergies.com/en/legal-notices/ - CFTC 2024 enforcement: https://www.cftc.gov/PressRoom/PressReleases/8953-24 - TotalEnergies annual reports: https://totalenergies.com/investors/publications-and-regulated-information/regulated-information/annual-financial-reports Historical priority: In 2024, the CFTC ordered TOTSA to pay $48 million for attempted manipulation of EBOB-linked futures. The order found that in March 2018 TOTSA sold large volumes of physical EBOB at below indicated bids while holding a large short EBOB-linked futures position, and that potentially relevant WhatsApp evidence was unavailable because communications were not timely preserved or produced. Preserve Commissioner Caroline Pham’s dissent and TOTSA’s defense/counterevidence. This prior adjudicated settlement is material background, but it does not establish that TOTSA’s 2026 crude trades were improper. ## 2026 non-futures and physical-market corruption/enforcement branch ### Iranian oil shipping, shell companies, and sanctions evasion Create a dedicated network around **Mohammad Hossein Shamkhani** and recurse into each person named by DOJ/Treasury in 2026 actions. 2026 root events: - March 6 DOJ civil-forfeiture complaints against more than $15.3 million allegedly connected to an Iranian oil distribution and shipping network; - April 15 Treasury sanctions against individuals, entities, and vessels in the Shamkhani network; - July 14 Treasury action targeting more than 50 additional individuals, entities, and vessels. Primary sources: - DOJ: https://www.justice.gov/opa/pr/united-states-files-civil-forfeiture-complaints-against-15m-funds-allegedly-linked-iranian - Treasury, April 15: https://home.treasury.gov/news/press-releases/sb0443 - Treasury, July 14: https://home.treasury.gov/news/press-releases/sb0562 Initial people include: - Mohammad Hossein Shamkhani; - Hossein Ghorbani Zahed; - Mohammad Reza Rahbar Madani; - Ali Rakhbarmadani; - Martin Austin Kaalund; - Alessandra Ronco; - Asghar Aghili Dehkordi; - Jijin George; - Behzad Moghadas; - Gautam Vishavdeep; - Maksim Tsernosjov; - every additional individual named in the 2026 complaints/designations. For each, recover corporate directorships, beneficial ownership, passports/nationalities only as officially published, shipping and commodities roles, vessel/company edges, sanctions history, litigation, prior employers, Russian-oil price-cap activity, and chronological changes. Treat DOJ complaint allegations as allegations and OFAC designations as official administrative actions, not criminal convictions. ### Ali Ansari and shadow exchange houses Create the July 10, 2026 Treasury event and recursively map **Ali Ansari**, named exchange houses, banks, shell companies, overseas assets, and beneficiaries. Treasury alleged that Ansari institutionalized large-scale embezzlement and used a global asset network to benefit Iranian regime elites. Source: - https://home.treasury.gov/news/press-releases/sb0558 ### Adani Enterprises Iranian LPG settlement Create the May 18, 2026 OFAC settlement event: - Adani Enterprises Limited agreed to pay **$275 million** to resolve potential civil liability for 32 apparent Iran-sanctions violations; - OFAC said the conduct was egregious and not voluntarily self-disclosed; - transactions involved LPG represented as Omani or Iraqi but allegedly originating in Iran and about $192.1 million in U.S.-dollar payments. Source: - https://ofac.treasury.gov/recent-actions/20260518 Recurse into the exact business unit, procurement/trading decision chain, compliance officers, Dubai supplier, shipping companies, vessels, banks, counterparties, investigators, and remedial actions. Do not assign personal culpability to Gautam Adani or any executive without evidence in the settlement or related records. ### Fuel procurement fraud Create 2026 conviction and sentencing events for **Jasen Butler**, owner of Independent Marine Oil Services LLC: - January 16 jury conviction on 34 felony counts; - April 8 sentence of 60 months; - more than $4.5 million in falsified fuel-related expenses submitted through military vessel procurement. Sources: - https://www.justice.gov/opa/pr/jury-convicts-florida-fuel-supplier-34-felonies-trial-multimillion-dollar-scheme-defraud-us - https://www.justice.gov/opa/pr/fuel-executive-gets-five-year-prison-sentence-defrauding-us-military-contract-bid-scam Deep-profile Butler, his company, employees, counterparties, SEA Card contracting chain, vessels, ports, invoices, shell entities, assets, prior businesses, and related litigation. ### Biofuel-credit fraud Create the May 29, 2026 sentencing event for **Christopher Burdett**, including General Manager **Royce Gillham**, the company, fraudulent Renewable Identification Numbers/credits, overstated production, tax-credit claims, restitution, and prior case history. Source: - https://www.justice.gov/opa/pr/florida-man-sentenced-18-months-prison-biofuel-fraud-conspiracy ### Oil-and-gas securities and offering cases Create separate 2026 enforcement-event records, preserving that many underlying schemes began before 2026: - Kevin N. Richards — April 7 final judgment involving alleged unregistered oil-and-gas offerings and undisclosed conflicts; - David P. Ortiz / DaveGlo Investment Group — April 27 final judgments; - Michael Bowen / Cannon Operating — June 8 final consent judgment; - additional SEC oil-and-gas offering cases filed, adjudicated, or settled in 2026. Sources: - https://www.sec.gov/enforcement-litigation/litigation-releases/lr-26531 - https://www.sec.gov/enforcement-litigation/litigation-releases/lr-26549 - https://www.sec.gov/enforcement-litigation/litigation-releases/lr-26575 Keep investor-offering fraud separate from commodity-price manipulation. ## Required recursive execution Run a root pass plus **six explicit recursive passes**. Every pass must create a `research-pass` record documenting inputs, queries, sources, records created, selected/deferred targets, confidence, unresolved identities, contradictions, and next frontier. ### Pass 0 — 2026 event ledger and market timeline - create one event per suspicious trading window, investigation, enforcement action, sanction, conviction, sentence, settlement, or forfeiture filing; - normalize every timestamp to UTC while preserving original timezone; - record announcement creation time, publication time, wire/headline distribution time, social-post time, exchange timestamp, market reaction, and source retrieval time separately; - obtain contract symbols, months, lot counts, notional estimates, prices, volume baselines, open interest, market depth, aggressor side, and related-product trades; - distinguish observed trade facts from profit estimates and legal allegations. ### Pass 1 — Account, firm, and execution-chain attribution - identify clearing members, FCMs, exchange members, brokers, market makers, funds, desks, strategy owners, and Tag 50/account identifiers from public records, court filings, regulatory requests, or later disclosures; - map Qube, Forza/Metabit, TOTSA, and all context firms; - distinguish automated headline response from directional discretionary trading; - compare suspicious-window activity against each firm’s ordinary baseline; - identify related positions in options, swaps, ETFs, equities, freight, crack spreads, calendar spreads, gasoline, diesel, currencies, prediction markets, and physical cargoes; - never infer the ultimate beneficial trader solely from an exchange member or clearing firm. ### Pass 2 — Deep person histories For every material person selected from Passes 0–1: - create a complete professional-role timeline; - map legal entities, funds, desks, boards, partnerships, ownership, signing authority, and compliance responsibility; - retrieve regulatory registrations, disciplinary history, litigation, sanctions, prior settlements, testimony, political/government roles, and material interlocks; - identify exact evidence linking or not linking the person to the 2026 event; - add an explicit `no-direct-evidence` or equivalent sourced claim when a person is included only because of governance position; - do not collect private addresses, personal phone numbers, family details, or unrelated personal information. ### Pass 3 — Government-information access chain For each March/April announcement: - reconstruct who drafted, reviewed, approved, scheduled, transmitted, translated, briefed, or technically published the decision; - use official schedules, visitor logs where public, congressional testimony, press-pool reports, archived posts, FOIA releases, ethics records, and staff biographies; - map Donald Trump and Abbas Araghchi as announcement/decision nodes; - identify U.S. and Iranian officials only where evidence places them in the relevant decision process; - distinguish access opportunity from evidence of a leak; - search for communications between government personnel and traders, donors, family businesses, lobbyists, funds, energy companies, or media only through lawful public records. ### Pass 4 — Physical benchmarks, cargoes, and energy-trading history - recurse from TOTSA’s prior EBOB manipulation settlement into responsible desks, people, benchmarks, brokers, Argus methodology, and controls; - inspect 2026 Platts Dubai/Oman/Murban pricing-window concentration, cargo trading, benchmark participation, and physical/derivative alignment without labeling concentrated legitimate trading as corruption absent evidence; - map comparable adjudicated energy-market cases involving Trafigura, Vitol, Glencore, BP, Shell, TotalEnergies, and others only as historical profiles attached to 2026 entities or people; - separate Platts, Argus, exchange, OTC, physical-cargo, freight, storage, and derivatives mechanisms. ### Pass 5 — Sanctions, procurement, credits, and offering fraud - recursively expand Shamkhani, Ansari, Adani LPG, Butler, Burdett, Richards, Ortiz, Bowen, and newly discovered 2026 cases; - map beneficial owners, directors, operators, traders, compliance personnel, counterparties, shell companies, vessels, banks, contracts, invoices, and proceeds; - distinguish sanctions designation, civil complaint, settlement, conviction, sentence, and allegation; - create cross-dataset links where people or organizations overlap Iran, Russia, Trump, defense procurement, TotalEnergies, shipping, commodities, or existing corruption datasets. ### Pass 6 — Regulators, surveillance gaps, and accountability - profile CFTC Chairman Michael S. Selig, enforcement leadership, relevant commissioners, CME/ICE surveillance leadership, DOJ prosecutors, congressional investigators, White House Counsel/ethics officials, and inspectors general; - obtain the CFTC response to Congress and current probe status; - assess 2026 staffing cuts, reporting delays, no-action letters, large-trader reporting changes, physical commodity swaps reporting, 24/7 crude-futures proposals, and enforcement capacity as regulatory context; - do not claim a rule change caused or enabled the trades without evidence; - map conflicts, recusals, revolving-door employment, lobbying, and political contributions only from authoritative filings; - preserve non-enforcement explanations and closed investigations. ## Canonical target selection After each pass, run the repository target selector with a pass-specific query: ```bash python3 scripts/starintel.py select-targets \ --query "2026 oil futures corruption market manipulation" \ --limit 20 \ --emit-documents \ --output recursive-targets-pass-N.jsonl ``` Selection priority: 1. person with documented control, decision access, trade authority, beneficial ownership, or enforcement relevance; 2. firm/account linked to a suspicious transaction cluster; 3. direct counterparty, broker, exchange, FCM, price-reporting agency, cargo, vessel, bank, or shell-company edge; 4. adjudicated prior conduct materially relevant to a 2026 entity/person; 5. high-value unresolved identity with a concrete primary-source acquisition path. Do not select a person merely because they work for a large firm. ## Required document classes Use exact dtypes exposed by `starintel_doc`; at minimum represent: - dataset-manifest; - research-pass; - event; - org; - person; - source; - claim; - relation; - investigation target/unresolved endpoint; - financial/market observation using the closest declared schema; - vessel, contract, product/system, or asset using exact available dtypes. ## Required outputs - `README.md`; - `sources.md`; - `manifest.json`; - canonical `starintel-documents.jsonl`; - Pass 0 through Pass 6 `research-pass` records; - `recursive-targets-pass-1.jsonl` through `recursive-targets-pass-6.jsonl`; - 2026 corruption-event ledger; - millisecond/second-resolution suspicious-trade timeline where data permits; - firm/account/clearing/exchange attribution matrix; - deep person dossiers and role timelines; - physical benchmark and cargo map; - sanctions/shipping/shell-company network; - enforcement-status matrix; - source-to-claim coverage table; - contradictions and alternative-explanations table; - unresolved beneficial-owner/account queue; - explicit next-pass targets. ## Evidence and accusation rules - The actual insiders, if any, are **not publicly identified** at dataset creation time. - A profitable or well-timed trade is not proof of insider trading. - Being named as a firm of interest is not an accusation or finding. - An algorithm reacting to a public headline is a viable alternative hypothesis and must be tested with timestamped data. - Market-maker gross or adjusted gains may not reflect the complete portfolio, hedges, spreads, or related-asset positions. - Governance title does not establish knowledge of, approval of, or participation in a trade. - Preserve government, company, commissioner, defendant, and defense explanations alongside allegations. - Label Senate statements as attributed political/oversight claims, not adjudicated facts. - Preserve exact legal posture: inquiry, investigation, request, allegation, complaint, designation, settlement, conviction, sentence, appeal, or closed matter. - Use public professional and corporate records only. No doxxing or private-person data. ## Validation and Git flow - Start from `main`. - Create branch `agent/oil-futures-2026-corruption`. - Inspect executable schemas before drafting documents: ```bash python3 scripts/starintel.py types python3 scripts/starintel.py schema --dtype <dtype> ``` - Do not hand-write normalized records under `db/`. - Write through `scripts/create-db-document.py` or import validated JSONL through `scripts/starintel.py import`. - Run: ```bash python3 scripts/validate-for-merge.py --site ``` - Open a draft PR only after the local gate succeeds. - Keep it draft until every required check passes on the current head. - Never weaken validation, hide unsupported claims in extensions, or merge invalid documents.

nsaspy added the

github-mirror

label

2026-09-08 17:05:44 +00:00

No labels

github-mirror

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

starintel-labs/starintel-gpt-auto-dig#997

No description provided.

Rows
Columns

[GH-#99] Create oil-futures: 2026 oil-market corruption events and deep person histories #997

Objective

Dataset

Hard temporal rule

Central 2026 suspicious-trading cluster

March 23, 2026

April 7, 2026

April 17, 2026

April 21, 2026

Combined scope

Additional 2026 suspicious oil-trading event

May 6, 2026

Congressional and executive-branch response

Warren–Whitehouse request

White House staff warning

Priority organizations and people for deep historical profiles

Qube Research & Technologies

Forza Fund Ltd. and Metabit Trading

TOTSA TotalEnergies Trading SA

2026 non-futures and physical-market corruption/enforcement branch

Iranian oil shipping, shell companies, and sanctions evasion

Ali Ansari and shadow exchange houses

Adani Enterprises Iranian LPG settlement

Fuel procurement fraud

Biofuel-credit fraud

Oil-and-gas securities and offering cases

Required recursive execution

Pass 0 — 2026 event ledger and market timeline

Pass 1 — Account, firm, and execution-chain attribution

Pass 2 — Deep person histories

Pass 3 — Government-information access chain

Pass 4 — Physical benchmarks, cargoes, and energy-trading history

Pass 5 — Sanctions, procurement, credits, and offering fraud

Pass 6 — Regulators, surveillance gaps, and accountability

Canonical target selection

Required document classes

Required outputs

Evidence and accusation rules

Validation and Git flow

[GH-#99] Create `oil-futures`: 2026 oil-market corruption events and deep person histories #997