Terrance DeJesus
45e6b901a2
* add description to hunting schema; change queries to be a list
* update createremotethreat by process hunt
* update dll hijack and masquerading as MSFT library
* remove sysmon specific dDLL hijack via masquerading MSFT library
* updated Masquerading Attempts as Native Windows Binaries
* updates Rare DLL Side-Loading by Occurrence
* updates Rare LSASS Process Access Attempts
* update DNS Queries via LOLBins with Low Occurence Frequency
* updated Low Occurrence of Drivers Loaded on Unique Hosts
* updates Excessive RDP Network Activity by Host and User
* updates Excessive SMB Network Activity by Process ID
* updated Executable File Creation by an Unusual Microsoft Binary
* Frequency of Process Execution and Network Logon by Source Address
* updates Frequency of Process Execution and Network Logon by Source Address
* updated Execution via Remote Services by Client Address
* updated Startup Execution with Low Occurrence Frequency by Unique Host
* updated Low Frequency of Process Execution via WMI by Unique Agent
* updated Low Frequency of Process Execution via Windows Scheduled Task by Unique Agent
* updated Low Occurence of Process Execution via Windows Services with Unique Agent
* Updated High Count of Network Connection Over Extended Period by Process
* update Libraries Loaded by svchost with Low Occurrence Frequency
* updated Microsoft Office Child Processes with Low Occurrence Frequency by Unique Agent
* updated Network Discovery via Sensitive Ports by Unusual Process
* updated PE File Transfer via SMB_Admin Shares by Agent or User
* updated Persistence via Run Key with Low Occurrence Frequency
* updates Persistence via Startup with Low Occurrence Frequency by Unique Host
* updates "Persistence via Run Key with Low Occurrence Frequency"; adjusted file names to remove data source
* updates "Low Occurrence of Suspicious Launch Agent or Launch Daemon"
* updates "Egress Network Connections with Total Bytes Greater than Threshold"
* updates "Rundll32 Execution Aggregated by Command Line"
* updates "Scheduled tasks Creation by Action via Registry"
* updates "Scheduled Tasks Creation for Unique Hosts by Task Command"
* updates "Suspicious Base64 Encoded Powershell Command"
* updates "Suspicious DNS TXT Record Lookups by Process"
* updates "Unique Windows Services Creation by Service File Name"
* Updates "Unique Windows Services Creation by Service File Name"
* updates "Windows Command and Scripting Interpreter from Unusual Parent Process"
* updates "Windows Logon Activity by Source IP"
* updates "Suspicious Network Connections by Unsigned Mach-O"
* updates LLM hunting queries
* re-generated markdown files; updated generate markdown py file
* updated test_hunt_data
* Update hunting/macos/queries/suspicious_network_connections_by_unsigned_macho.toml
Co-authored-by: Mika Ayenson <Mikaayenson@users.noreply.github.com>
* Update hunting/windows/queries/drivers_load_with_low_occurrence_frequency.toml
Co-authored-by: Mika Ayenson <Mikaayenson@users.noreply.github.com>
* Update hunting/windows/queries/domain_names_queried_via_lolbins_and_with_low_occurence_frequency.toml
Co-authored-by: Mika Ayenson <Mikaayenson@users.noreply.github.com>
* Update hunting/windows/queries/excessive_rdp_network_activity_by_source_host_and_user.toml
Co-authored-by: Mika Ayenson <Mikaayenson@users.noreply.github.com>
* Update hunting/windows/queries/excessive_rdp_network_activity_by_source_host_and_user.toml
Co-authored-by: Mika Ayenson <Mikaayenson@users.noreply.github.com>
* updated missing integrations
* updated MD docs according to recent hunting changes
* Update hunting/windows/queries/executable_file_creation_by_an_unusual_microsoft_binary.toml
Co-authored-by: Jonhnathan <26856693+w0rk3r@users.noreply.github.com>
* Update hunting/windows/queries/detect_rare_dll_sideload_by_occurrence.toml
Co-authored-by: Jonhnathan <26856693+w0rk3r@users.noreply.github.com>
* Update hunting/windows/queries/detect_masquerading_attempts_as_native_windows_binaries.toml
Co-authored-by: Jonhnathan <26856693+w0rk3r@users.noreply.github.com>
* Update hunting/windows/queries/detect_dll_hijack_via_masquerading_as_microsoft_native_libraries.toml
Co-authored-by: Jonhnathan <26856693+w0rk3r@users.noreply.github.com>
* Update hunting/llm/queries/aws_bedrock_dos_resource_exhaustion_detection.toml
Co-authored-by: Jonhnathan <26856693+w0rk3r@users.noreply.github.com>
* added enrichment policy link to rule
* Update hunting/windows/docs/execution_via_windows_management_instrumentation_by_occurrence_frequency_by_unique_agent.md
Co-authored-by: Samirbous <64742097+Samirbous@users.noreply.github.com>
* Update hunting/windows/docs/windows_command_and_scripting_interpreter_from_unusual_parent.md
Co-authored-by: Samirbous <64742097+Samirbous@users.noreply.github.com>
* Update hunting/windows/docs/windows_command_and_scripting_interpreter_from_unusual_parent.md
Co-authored-by: Samirbous <64742097+Samirbous@users.noreply.github.com>
* Update hunting/windows/docs/rundll32_execution_aggregated_by_cmdline.md
Co-authored-by: Samirbous <64742097+Samirbous@users.noreply.github.com>
* Update hunting/windows/docs/microsoft_office_child_processes_with_low_occurrence_frequency.md
Co-authored-by: Samirbous <64742097+Samirbous@users.noreply.github.com>
* Update hunting/windows/docs/microsoft_office_child_processes_with_low_occurrence_frequency.md
Co-authored-by: Samirbous <64742097+Samirbous@users.noreply.github.com>
* Update hunting/windows/queries/execution_via_windows_management_instrumentation_by_occurrence_frequency_by_unique_agent.toml
Co-authored-by: Samirbous <64742097+Samirbous@users.noreply.github.com>
* Update hunting/windows/queries/execution_via_windows_management_instrumentation_by_occurrence_frequency_by_unique_agent.toml
Co-authored-by: Samirbous <64742097+Samirbous@users.noreply.github.com>
* Update hunting/index.md
Co-authored-by: Samirbous <64742097+Samirbous@users.noreply.github.com>
* Update hunting/windows/docs/execution_via_network_logon_by_occurrence_frequency_by_top_source_ip.md
Co-authored-by: Samirbous <64742097+Samirbous@users.noreply.github.com>
* Update hunting/windows/queries/execution_via_network_logon_by_occurrence_frequency_by_top_source_ip.toml
Co-authored-by: Samirbous <64742097+Samirbous@users.noreply.github.com>
---------
Co-authored-by: Mika Ayenson <Mikaayenson@users.noreply.github.com>
Co-authored-by: Jonhnathan <26856693+w0rk3r@users.noreply.github.com>
Co-authored-by: Samirbous <64742097+Samirbous@users.noreply.github.com>
(cherry picked from commit 632e169f7a)
LLM Threat Hunting Queries
Welcome to the LLM subfolder within the hunting directory of the detection-rules repository. This specialized section is dedicated to threat hunting queries designed for Large Language Model (LLM) applications, targeting the unique security challenges these systems face.
Emphasis on OWASP Top 10 for LLMs
Our queries are developed with a keen awareness of the OWASP Top 10 risks for Large Language Model Applications. This crucial resource outlines the predominant security risks for LLMs, guiding our efforts in crafting queries that proactively address these vulnerabilities and ensure comprehensive threat mitigation.
Emphasis on MITRE ATLAS
The ATLAS Matrix covers the progression of tactics used in attacks with ML techniques belonging to different tactics.
- Reconnaissance
- Resource Development
- Initial Access
- ML Model Access
- Execution
- Persistence
- Privilege Escalation
- Defense Evasion
- Credential Access
- Discovery
- Collection
- ML Attack Staging
- Exfiltration
- Impact
Scope of Threats and Protections
The queries in this folder are tailored to monitor and protect against a broad spectrum of threats to LLMs:
- Sensitive Content Refusal: Monitors LLM interactions to ensure compliance with ethical standards, particularly in refusing to process sensitive topics.
- Denial of Service (DoS) and Resource Exhaustion: Aims to prevent disruptions in LLM operations by detecting patterns indicative of DoS attacks or resource exhaustion scenarios.
- Latency Anomalies: Tracks processing delays that could signal underlying performance issues or security threats, maintaining operational efficiency and safeguarding against potential attacks like DDoS.
Benefits of These Queries
These queries assist organizations in:
- Detecting and mitigating misuse or attacks that threaten data integrity or disrupt services.
- Ensuring that LLMs adhere strictly to operational and ethical boundaries through continuous monitoring.
- Maintaining high performance and reliability of LLMs by preemptively identifying and resolving factors that cause inefficiencies.
For more details, read our blog on LLM Detections.