{"schema_version":"0.8.0","id":"LR-2026-0239","record_version":2,"title":"BUGSTER uses DeepSeek-powered Hermes Agent to breach university MongoDB, stealing 5,200+ records","verification":{"human_verified":false,"method":"automated_gate","auto_verified":true,"score":66.4,"score_breakdown":{"source_authority":12.1,"independent_corroboration":15.0,"implicated_party_confirmed":0.0,"citation_integrity":14.0,"specificity":9.0,"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},"gate_version":"gate-2026-10-18.7","cross_check_model":null,"cross_check_agreement":null,"duplicate_of":null,"verified_by":null,"verified_at":"2026-10-23","extraction_run_id":null},"system":{"model_name":"DeepSeek V4","model_version":null,"developer":"DeepSeek","harness":"Hermes Agent","tools_granted":["credential_scanning","password_cracking","coding","code generation (Cursor)","Telegram bot assistant profiles","brute-force tooling","Wi-Fi capture and cloud GPU cracking","Discord webhooks"],"permissions_scope":"Attacker-configured environment with local permission and sandbox controls disabled, intended to reduce restrictions on malicious tasks.","autonomy_level":"unknown","guardrails_present":[],"deployment_context":"production","tier":"enterprise","lifecycle_stage":"production","integration_type":"agentic_with_tools","intended_use":"AI assistant framework used to support attacks on Indian organizations, including mapping targets, writing scripts and reports","safeguard_state":"safety controls turned off","production_intended":null},"event":{"date_occurred":null,"date_disclosed":"2026-10-23","incident_class":"security_compromise","failure_class":null,"agent_involved":true,"secondary_failure_classes":[],"trigger":"human_misuse","duration":null,"detection_method":"third_party","summary":"In one confirmed intrusion, the attacker stole a MongoDB database containing more than 5,200 records, including personal information about students and staff.","tags":["ai-assisted-attack","credential-theft","database-exfiltration","pii-theft","education-sector","jailbreak-configuration"],"date_detected":null,"near_miss":null,"root_cause":"integration","root_cause_detail":"The report attributes the intrusions to weak administrator account security and exposed services rather than a software flaw.","failure_class_detail":null,"responsible_party":"third_party","adversarial_actor":"external_attacker","input_provenance":null,"time_to_contain":null,"operations_paused":null,"remediation":"Review administrator accounts, remove weak or reused passwords, require multi-factor authentication, limit database access, check internet-facing portals, monitor account activity, keep secrets out of public files and logs, revoke or rotate exposed credentials, and limit AI agent permissions with safety and approval controls enabled.","precursor_events":null},"impact":{"direct_loss_usd":null,"remediation_hours":null,"records_exposed":null,"third_parties_affected_count":null,"severity":"moderate","severity_because":"moderate, because data or systems of others touched, with a bounded scope stated: Personal data of students and staff at two universities was stolen, a bounded scope of more than 5,200 records.","harm_class":["privacy","third_party_security"],"affected_description":"two unnamed Indian universities; MongoDB database with student and faculty records","protected_class_involved":null,"data_categories_exposed":["pii","credentials"],"loss_range_usd":null,"reversibility":null,"downstream_exposure_count":null},"evidence":{"sources":[{"url":"https://cybersecuritynews.com/hacker-with-limited-skills-used-deepseek-ai/amp/","publisher":"news search (Google News index)","date":"2026-10-23","type":"press","title":"Hacker With Limited Skills Used DeepSeek AI to Breach Indian University and Steal Over 5,200 Records","archived_url":null,"attachment":"first_report"},{"url":"https://cyberpress.org/ai-assisted-hacker-breaches-university/","publisher":"news search (Google News index)","date":"2026-10-23","type":"press","title":"Low-Skilled Hacker Used AI Agents to Hack University and Steal Thousands of Records","archived_url":null,"attachment":"fingerprint"}],"per_field_confidence":{"system.model_name":"C","system.harness":"C","system.tools_granted":"C","system.permissions_scope":"C","system.tier":"D","system.lifecycle_stage":"D","system.integration_type":"D","system.intended_use":"C","system.safeguard_state":"C","system.developer":"C","system.deployment_context":"D","system.autonomy_level":"D","event.tags":"D","event.root_cause":"D","event.responsible_party":"D","event.adversarial_actor":"C","event.remediation":"C","event.agent_involved":"C","event.trigger":"C","event.detection_method":"D","event.summary":"C","event.incident_class":"C","event.date_disclosed":"A","impact.harm_class":"C","impact.affected_description":"C","impact.data_categories_exposed":"C","impact.severity":"D","context.industry":"C","context.denominator_available":"D","context.country":"C","event.root_cause_detail":"D","impact.severity_because":"D"},"field_citations":{"system.model_name":[{"source_index":0,"quote":"The actor’s setup combined Hermes Agent, an AI assistant framework, with DeepSeek V4 as its reasoning model and Cursor as a coding tool."}],"system.developer":[{"source_index":0,"quote":"The actor’s setup combined Hermes Agent, an AI assistant framework, with DeepSeek V4 as its reasoning model and Cursor as a coding tool."}],"system.harness":[{"source_index":0,"quote":"The actor’s setup combined Hermes Agent, an AI assistant framework, with DeepSeek V4 as its reasoning model and Cursor as a coding tool."}],"system.tools_granted":[{"source_index":0,"quote":"Investigators also found a large credential-scanning toolkit designed to look for exposed secrets, such as cloud keys and private keys in files or logs."},{"source_index":1,"quote":"Telegram bots provided access to specialized assistant profiles, while the Cursor development environment supported code generation."}],"system.permissions_scope":[{"source_index":1,"quote":"The actor configured prompt overrides and disabled local permission and sandbox controls, creating an environment intended to reduce restrictions on malicious tasks."}],"system.safeguard_state":[{"source_index":0,"quote":"Unit 42 said the attacker turned off safety controls and relied on AI-written scripts and reports."}],"system.deployment_context":[{"source_index":1,"quote":"At a second Indian public university, the actor obtained an administrator account protected by a weak password and exfiltrated a MongoDB database."}],"system.lifecycle_stage":[{"source_index":1,"quote":"At a second Indian public university, the actor obtained an administrator account protected by a weak password and exfiltrated a MongoDB database."}],"system.integration_type":[{"source_index":1,"quote":"BUGSTER used the Hermes Agent framework with DeepSeek V4 as its main reasoning engine."}],"system.intended_use":[{"source_index":0,"quote":"Unit 42 said the actor used AI-generated reports to organize research and produce scripts."}],"event.trigger":[{"source_index":0,"quote":"A cybercriminal using the online handle BUGSTER used artificial intelligence tools to support attacks on Indian organizations, including two universities."}],"event.detection_method":[{"source_index":1,"quote":"A low-skilled hacker used artificial intelligence assistants to attack Indian universities and steal more than 5,200 records containing personal information, according to Palo Alto Networks’ Unit 42."}],"event.summary":[{"source_index":0,"quote":"In one confirmed intrusion, the attacker stole a MongoDB database containing more than 5,200 records, including personal information about students and staff, according to Palo Alto Networks’ Unit 42."}],"event.tags":[{"source_index":1,"quote":"Researchers found evidence of credential theft, database exfiltration, telecommunications fraud planning, Wi-Fi password cracking and cryptocurrency reconnaissance."}],"event.root_cause":[{"source_index":0,"quote":"In the second case, the actor accessed a public university’s database with an administrator account protected by a weak password."}],"event.root_cause_detail":[{"source_index":0,"quote":"These incidents show why an exposed service or weak account can have serious effects, even without a newly discovered software flaw."}],"event.responsible_party":[{"source_index":1,"quote":"A low-skilled hacker used artificial intelligence assistants to attack Indian universities and steal more than 5,200 records containing personal information, according to Palo Alto Networks’ Unit 42."}],"event.adversarial_actor":[{"source_index":0,"quote":"A cybercriminal using the online handle BUGSTER used artificial intelligence tools to support attacks on Indian organizations, including two universities."}],"event.incident_class":[{"source_index":0,"quote":"In one confirmed intrusion, the attacker stole a MongoDB database containing more than 5,200 records, including personal information about students and staff, according to Palo Alto Networks’ Unit 42."}],"event.agent_involved":[{"source_index":0,"quote":"The actor’s setup combined Hermes Agent, an AI assistant framework, with DeepSeek V4 as its reasoning model and Cursor as a coding tool."}],"event.remediation":[{"source_index":0,"quote":"Universities and other organizations should review administrator accounts, remove weak or reused passwords, and require multi-factor authentication where possible."}],"impact.harm_class":[{"source_index":0,"quote":"In one confirmed intrusion, the attacker stole a MongoDB database containing more than 5,200 records, including personal information about students and staff, according to Palo Alto Networks’ Unit 42."}],"impact.affected_description":[{"source_index":0,"quote":"The stolen MongoDB data included student records from diploma, undergraduate and postgraduate programs, as well as faculty information."}],"impact.data_categories_exposed":[{"source_index":0,"quote":"In one confirmed intrusion, the attacker stole a MongoDB database containing more than 5,200 records, including personal information about students and staff, according to Palo Alto Networks’ Unit 42."},{"source_index":1,"quote":"The actor separately examined an Indian e-commerce website, collecting a storefront API token, administrator email addresses, phone numbers and payment merchant identifiers."}],"context.industry":[{"source_index":0,"quote":"A cybercriminal using the online handle BUGSTER used artificial intelligence tools to support attacks on Indian organizations, including two universities."}],"context.country":[{"source_index":0,"quote":"A cybercriminal using the online handle BUGSTER used artificial intelligence tools to support attacks on Indian organizations, including two universities."}],"impact.severity":[{"source_index":0,"quote":"In one confirmed intrusion, the attacker stole a MongoDB database containing more than 5,200 records, including personal information about students and staff, according to Palo Alto Networks’ Unit 42."}],"impact.severity_because":[{"source_index":0,"quote":"In one confirmed intrusion, the attacker stole a MongoDB database containing more than 5,200 records, including personal information about students and staff, according to Palo Alto Networks’ Unit 42."}],"event.date_disclosed":[{"source_index":0,"quote":"published: 2026-10-23"}]},"notes_on_gaps":"No source states event.date_occurred, date_detected, duration, time_to_contain, operations_paused, precursor_events, input_provenance, direct_loss_usd, remediation_hours, third_parties_affected_count, protected_class_involved, reversibility, downstream_exposure_count, org_size_band, denominator, systemic_exposure, shared_model_id, upstream_incident_id, incident_evidence, regulatory or coverage. event.failure_class left null: the sources describe an attacker using the agent, not an agent-mechanism failure. system.autonomy_level left unknown: no source states the human-in-the-loop arrangement. system.guardrails_present left empty: sources state controls were disabled by the attacker, not which controls existed. system.tier not written (derived). Severity: source 0 says moderate (bounded 5,200 records), source 1 says major (many third parties across multiple universities); I took moderate as the better-supported bounded-scope value and recorded the conflict. event.root_cause 'integration' is a grade-D classification from the weak-password administrator account sentence; the source frames it as weak account security and exposed services, not a software flaw. event.detection_method 'third_party' rests on Unit 42 (a vendor) reporting the activity; no source states the first signal. event.trigger 'human_misuse' rests on the attacker deliberately using the tools. No source states containment, notification or a control actually changed; remediation is only recommended. impact.records_exposed: the source gives 5200 only as a bound ('more than'); not stored as an exact count. system.tier: enterprise by rule (operator BUGSTER (attacker); target universities unnamed differs from developer DeepSeek). Sources disagree on impact.affected_description: [0] \"two unnamed Indian universities; MongoDB database with student and faculty records\", [1] \"Two Indian universities (one online education platform, one public university) and an Indian e-commerce website; student and faculty records.\"; impact.severity: [0] \"moderate\", [1] \"major\"; system.harness: [0] \"Hermes Agent\", [1] \"Hermes Agent framework\"; system.intended_use: [0] \"AI assistant framework used to support attacks on Indian organizations, including mapping targets, writing scripts and reports\", [1] \"Attacker's reasoning engine for reconnaissance, script generation and conversational guidance during intrusions.\"; system.safeguard_state: [0] \"safety controls turned off\", [1] \"Prompt overrides configured and local permission and sandbox controls disabled by the attacker.\". [aggregated by deepseek-chat, aggregate-2026-09-27.1, 2026-10-23] system.lifecycle_stage: the systems attacked were live production systems of the universities. system.integration_type: the model ran inside an agent framework with tooling (Cursor, Telegram bots, cracking tooling). system.deployment_context: classified from described behaviour. system.autonomy_level: no source states the arrangement. event.tags: classified from described behaviour. event.responsible_party: the defect is owned by an external attacker, not the provider, integrator or deployer. event.detection_method: a source states an outside organisation (a vendor, a partner, a platform) noticed it first. context.denominator_available: no denominator is available for this incident."},"context":{"industry":"Educational Services","org_size_band":null,"country":"IN","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":"BUGSTER ran the Hermes Agent framework with DeepSeek V4, plus Cursor and Telegram assistant profiles, to research targets, write scripts and reports, and scan for credentials against Indian universities.","quote":"The actor’s setup combined Hermes Agent, an AI assistant framework, with DeepSeek V4 as its reasoning model and Cursor as a coding tool.","source_index":0,"grade":"C"},{"step":"went_wrong","text":"The AI-assisted tooling was used to break into university systems, including taking an administrator account protected by a weak password, crossing the universities' access boundaries.","quote":"At a second Indian public university, the actor obtained an administrator account protected by a weak password and exfiltrated a MongoDB database.","source_index":1,"grade":"C"},{"step":"reached","text":"A MongoDB database with more than 5,200 student and staff records was stolen, and an e-commerce storefront API token, admin emails and payment merchant identifiers were collected.","quote":"In one confirmed intrusion, the attacker stole a MongoDB database containing more than 5,200 records, including personal information about students and staff, according to Palo Alto Networks’ Unit 42.","source_index":0,"grade":"C"},{"step":"caught","text":"Palo Alto Networks' Unit 42 researchers identified and investigated the activity by examining the operator's environment, tools and records.","quote":"Analysts from Unit 42 identified and investigated the activity by examining the operator’s environment, tools and records.","source_index":0,"grade":"C"},{"step":"done","text":"No containment or notification is stated; the report recommends reviewing administrator accounts, removing weak passwords, requiring multi-factor authentication, limiting database access and limiting AI agent permissions.","quote":"Universities and other organizations should review administrator accounts, remove weak or reused passwords, and require multi-factor authentication where possible.","source_index":0,"grade":"C"}],"model":"deepseek-chat","prompt_version":"aggregate-2026-09-27.1","extracted_at":"2026-10-23","source_urls":["https://cybersecuritynews.com/hacker-with-limited-skills-used-deepseek-ai/amp/","https://cyberpress.org/ai-assisted-hacker-breaches-university/"]}}