{"schema_version":"0.10.0","id":"LR-2026-0288","record_version":2,"title":"Fake Polymarket platforms use LLM chat agents with provider-switching to defraud victims","verification":{"human_verified":false,"method":"automated_gate","auto_verified":true,"score":48.5,"score_breakdown":{"source_authority":12.1,"independent_corroboration":7.0,"implicated_party_confirmed":0.0,"citation_integrity":14.0,"specificity":0.0,"completeness":8.0,"unhedged_language":5.0},"checks":{"schema_valid":true,"every_field_cited":true,"quotes_verbatim":true,"source_fetched":true,"distinct_incident":true,"dates_plausible":true,"sources_agree":true,"source_describes_incident":true,"severe_backed_by_source":true,"record_integrity":true,"eligible":true,"about_this_case":true},"gate_version":"gate-2026-10-18.7","cross_check_model":null,"cross_check_agreement":null,"duplicate_of":null,"verified_by":null,"verified_at":"2026-10-31","extraction_run_id":null,"evidence_level":{"level":"preliminary","reason":"1 report: Cyber Security News · subject to change","original_document":null,"reports_total":1,"reports_independent":1,"reports_derivative":0,"developing":true,"disputed":false,"rules_version":"evidence-2026-10-29.1","computed_at":"2026-10-31","history":[{"date":"2026-10-31","from_level":null,"to_level":"preliminary","reason":"1 report: Cyber Security News · subject to change"}]},"sources_hash":"d61a9b4b642b42f8907c8fb8cbddef8479ba79288941fa4a9bcb4a1a96fc4167"},"system":{"model_name":"unknown","model_version":null,"developer":"unknown","harness":"custom scaffold","tools_granted":[],"permissions_scope":null,"autonomy_level":"unknown","guardrails_present":[],"deployment_context":"production","tier":null,"lifecycle_stage":null,"integration_type":"embedded","intended_use":"Answer support messages and translate victim replies into Russian on fake prediction market sites.","safeguard_state":"If one model refuses a request, the system switches to a model from another provider.","production_intended":null},"event":{"date_occurred":null,"date_disclosed":"2026-10-30","incident_class":"security_compromise","failure_class":null,"agent_involved":false,"secondary_failure_classes":[],"trigger":"unknown","duration":null,"detection_method":"public_researcher","summary":"Three Russian-speaking fraud teams (AtlasCore, PolyFucked, Magnetic Team) sell ready-made fake Polymarket-style prediction market sites to affiliates.","tags":["fake-prediction-market","cryptocurrency-theft","withdrawal-blocking","affiliate-fraud","identity-document-harvesting","llm-support-chat"],"date_detected":null,"near_miss":false,"root_cause":null,"root_cause_detail":null,"failure_class_detail":null,"responsible_party":"third_party","adversarial_actor":"external_attacker","input_provenance":null,"time_to_contain":null,"operations_paused":null,"remediation":null,"precursor_events":null},"impact":{"direct_loss_usd":null,"remediation_hours":null,"records_exposed":null,"third_parties_affected_count":null,"severity":"major","severity_because":"major: Many victims across multiple fake platforms lost cryptocurrency, with self-reported deposits of about $10.97 million, though the figures are unverified.","harm_class":["financial","privacy"],"affected_description":"Victims who deposit cryptocurrency on fake Polymarket-style prediction market sites","protected_class_involved":null,"data_categories_exposed":["card numbers","CVV codes","passport or driver's license photos","selfie holding the document"],"loss_range_usd":null,"reversibility":null,"downstream_exposure_count":null},"evidence":{"sources":[{"url":"https://cybersecuritynews.com/fake-polymarket-platforms/","publisher":"Cyber Security News","date":"2026-10-30","type":"press","title":"Hackers Build Fake Polymarket Platforms That Rig Trades and Block Cryptocurrency Withdrawals","archived_url":null,"attachment":"first_report","independence":"independent"}],"per_field_confidence":{"system.model_name":"D","system.harness":"D","system.integration_type":"D","system.intended_use":"C","system.safeguard_state":"C","system.developer":"D","system.deployment_context":"D","system.autonomy_level":"D","event.tags":"D","event.near_miss":"D","event.responsible_party":"D","event.adversarial_actor":"C","event.agent_involved":"D","event.detection_method":"C","event.summary":"D","event.incident_class":"D","event.date_disclosed":"A","event.trigger":"D","impact.harm_class":"D","impact.affected_description":"C","impact.data_categories_exposed":"C","impact.severity":"D","context.industry":"D","context.denominator_available":"D","context.country":"D","impact.severity_because":"D"},"field_citations":{"system.model_name":[{"source_index":0,"quote":"If one model refuses a request, the system switches to a model from another provider."}],"system.developer":[{"source_index":0,"quote":"Code from the fake sites shows operators connecting a language model to answer support messages and translate victim replies into Russian."}],"system.harness":[{"source_index":0,"quote":"Code from the fake sites shows operators connecting a language model to answer support messages and translate victim replies into Russian."}],"system.integration_type":[{"source_index":0,"quote":"Code from the fake sites shows operators connecting a language model to answer support messages and translate victim replies into Russian."}],"system.intended_use":[{"source_index":0,"quote":"Code from the fake sites shows operators connecting a language model to answer support messages and translate victim replies into Russian."}],"system.safeguard_state":[{"source_index":0,"quote":"If one model refuses a request, the system switches to a model from another provider."}],"system.deployment_context":[{"source_index":0,"quote":"They use the copies to take cryptocurrency from people who think they are placing real bets."}],"system.autonomy_level":[{"source_index":0,"quote":"Code from the fake sites shows operators connecting a language model to answer support messages and translate victim replies into Russian."}],"event.detection_method":[{"source_index":0,"quote":"Researchers at Recorded Future’s Insikt Group identified the three teams and the tools they sell."}],"event.summary":[{"source_index":0,"quote":"Three Russian-speaking “traffer” teams, called AtlasCore, PolyFucked and Magnetic Team, sell ready-made fake prediction markets to affiliates."}],"event.tags":[{"source_index":0,"quote":"Code from the fake sites shows operators connecting a language model to answer support messages and translate victim replies into Russian."}],"event.near_miss":[{"source_index":0,"quote":"Once a victim deposits money, the site decides every win and loss, and withdrawals stop working."}],"event.adversarial_actor":[{"source_index":0,"quote":"Three Russian-speaking “traffer” teams, called AtlasCore, PolyFucked and Magnetic Team, sell ready-made fake prediction markets to affiliates."}],"event.incident_class":[{"source_index":0,"quote":"Criminals are copying the look of Polymarket, one of the biggest prediction market sites."}],"event.agent_involved":[{"source_index":0,"quote":"Code from the fake sites shows operators connecting a language model to answer support messages and translate victim replies into Russian."}],"event.responsible_party":[{"source_index":0,"quote":"Three Russian-speaking “traffer” teams, called AtlasCore, PolyFucked and Magnetic Team, sell ready-made fake prediction markets to affiliates."}],"impact.affected_description":[{"source_index":0,"quote":"They use the copies to take cryptocurrency from people who think they are placing real bets."}],"impact.data_categories_exposed":[{"source_index":0,"quote":"Victims are asked for card numbers, CVV codes, passport or driver’s license photos and a selfie holding the document."}],"impact.harm_class":[{"source_index":0,"quote":"They use the copies to take cryptocurrency from people who think they are placing real bets."}],"impact.severity_because":[{"source_index":0,"quote":"AtlasCore’s Telegram channel claimed about $10.97 million in deposits between May 2 and September 18, 2026."}],"context.industry":[{"source_index":0,"quote":"Criminals are copying the look of Polymarket, one of the biggest prediction market sites."}],"context.country":[{"source_index":0,"quote":"Three Russian-speaking “traffer” teams, called AtlasCore, PolyFucked and Magnetic Team, sell ready-made fake prediction markets to affiliates."}],"impact.severity":[{"source_index":0,"quote":"AtlasCore’s Telegram channel claimed about $10.97 million in deposits between May 2 and September 18, 2026."}],"event.date_disclosed":[{"source_index":0,"quote":"published: 2026-10-30"}]},"notes_on_gaps":"No source names the model family, version, developer, or the operator's country; the LLM is only described as connected by the site operators. No source states the agent's approval arrangement, so autonomy_level is unknown. No source states a cause for the LLM behaviour, so trigger, failure_class, root_cause and input_provenance are omitted. No source states a date the incident occurred or was disclosed, a loss figure in USD for victims (the $10.97M is the criminals' self-reported deposits, unverified), a victim count, or remediation. incident_class security_compromise is a judgement: the harm is fraud by human operators, not a breach of the model's own boundary; the LLM's role was support chat. agent_involved false because the LLM answered chats rather than acting with tools. severity major rests on the unverified self-reported deposit figure and many victims; the dollar band would otherwise be severe if the figure were verified. system.tier: not set; no operator named and deployment_context is not a lab's. [aggregated by deepseek-chat, aggregate-2026-09-27.1, 2026-10-31] system.model_name: classified from described behaviour. system.harness: classified from described behaviour. system.integration_type: classified from described behaviour. system.developer: no source names the developer. system.deployment_context: classified from described behaviour. system.autonomy_level: no source states the arrangement. event.tags: classified from described behaviour. event.near_miss: classified from described behaviour. event.responsible_party: classified from described behaviour. event.agent_involved: classified from described behaviour. event.summary: classified from described behaviour. event.incident_class: a source states an attacker, an injected instruction or a misused credential caused the system to act, or that the system's own boundary (sandbox, permissions, weights, accounts) was breached. event.trigger: no source states a cause. impact.harm_class: classified from described behaviour. context.industry: classified from described behaviour. context.denominator_available: no denominator is available for this incident. context.country: no source states the operator's location."},"context":{"industry":"Finance and Insurance","org_size_band":null,"country":"unknown","denominator_available":false,"denominator":null,"related_incident_ids":[],"systemic_exposure":null,"shared_model_id":null,"upstream_incident_id":null,"incident_evidence":null,"regulatory":null,"coverage":null},"narrative":{"steps":[{"step":"doing","text":"Fraud teams sold fake Polymarket-style prediction market sites to affiliates; the sites took crypto deposits, faked trading activity, and used a language model to answer victim support chats.","quote":"Code from the fake sites shows operators connecting a language model to answer support messages and translate victim replies into Russian.","source_index":0,"grade":"C"},{"step":"went_wrong","text":"The sites rigged every win and loss and blocked withdrawals behind fake fees and checks, taking victims' cryptocurrency.","quote":"Once a victim deposits money, the site decides every win and loss, and withdrawals stop working.","source_index":0,"grade":"C"},{"step":"reached","text":"Victims' crypto deposits and personal data, including card numbers, CVV codes and passport photos, were collected across many fake domains.","quote":"Victims are asked for card numbers, CVV codes, passport or driver’s license photos and a selfie holding the document.","source_index":0,"grade":"C"},{"step":"caught","text":"Recorded Future's Insikt Group researchers identified the three teams and the tools they sell.","quote":"Researchers at Recorded Future’s Insikt Group identified the three teams and the tools they sell.","source_index":0,"grade":"C"},{"step":"done","text":"The report names the teams and publishes indicators of compromise so security teams can block domains and IPs and watch for new lookalike domains.","quote":"Security teams can block the domains and IPs below and watch for new lookalike domains.","source_index":0,"grade":"C"}],"model":"deepseek-chat","prompt_version":"aggregate-2026-09-27.1","extracted_at":"2026-10-31","source_urls":["https://cybersecuritynews.com/fake-polymarket-platforms/"]}}