Stellar Cyber 6.6.0s Release Notes

Software Release Date:
Release Note Updated:

The Stellar Cyber 6.6.0s release delivers the following updates to the Stellar Cyber Open XDR platform.

The release notes are organized into the following sections:

Highlights

Autonomous SOC / Auto Triage

  • AI Assistant (Early Access Program): Added AI Assistant to every automatically triaged case so you can investigate cases and alerts in plain language, asking why a case received its verdict, clarifying alert details, and exploring follow-up hypotheses against the case data and Verdict Signal Check (VSC) context that Auto Triage has already assembled.

  • Auto Triage Verdict Visibility: Auto Triage verdicts now appear as filterable columns in the Alert Table and Threat Hunting views. In addition, a response action panel was added to the Auto Triage alert page (including phishing email alerts) so analysts can see and act on triage outcomes without opening individual cases.

System / Platform

  • Platform Health Monitoring in the System Action Center (Early Access Program): Centralized platform health monitoring alerts in the System Action Center for improved visibility and faster response to platform issues.

  • License Enforcement and Usage Notifications: Added API actions for license enforcement and usage notifications.

Detections/Machine Learning

  • Improved Login Anomaly Fidelity: Alert suppression for Impossible Travel Anomaly and User Login Location Anomaly is now customizable, Impossible Travel prioritizes records with usernames, and ASN enrichment fields were added to Impossible Travel Anomaly alerts.

  • Improved User Counting Accuracy: Improved the accuracy of license user counting by integrating external data sources for Microsoft Entra ID.

Integrations

Usability

  • Watchlists for All Alert Fields: Enabled Add to Watchlistfor all alert fields.

  • Selective Parser Port Activation: Parser ingestion ports can now be enabled on-demand and parsers added after 6.5.0 are inactive by default, reducing false alerts from unused listeners.

Actions Required

There are no actions required in this release.

Behavior Changes

Changes that affect the way users interact with the product or interpret results are listed below.

  • DATA-3412: The totalbytes field in the Fortinet FortiAnalyzer parser is now calculated as the sum of inbytes_total and outbytes_total, consistent with how other parsers calculate this field. Previously, the FortiAnalyzer parser calculated totalbytes as the sum of inbytes_delta and outbytes_delta, which significantly underreported session volume when delta values differed from totals. Dashboards, detections, or queries that rely on totalbytes from FortiAnalyzer data may return different values after this change.

  • AELDEV-71104: Built-in legacy parsers added in 6.5.0 and later releases are now inactive by default for new tenants and must be explicitly activated in Parser Studio before use. Previously added legacy parsers retain their existing active status and are not affected by this change. You can activate or deactivate any individual parser at any time from Parser Studio, and the configuration is applied to sensors on the next deployment.

  • AELDEV-64683: The Google Workspace log collector no longer maps the metadata.customerId field to user.id. The customerId value is a Google account identifier for the organization, not a user identifier, so this mapping was incorrect and caused confusion in alert context. Existing events already indexed with this normalization are not affected. Newly ingested Google Workspace events no longer populate user.id from this field.

Deprecated Features

There are no deprecated features in this release.

Autonomous SOC

Improvements

  • ASOC-142: Added Auto Triage verdict columns to the Alert Table and Threat Hunting views.

    Added Auto Triage verdicts as columns to both the Alert Table and Threat Hunting views, enabling visibility and filtering of verdicts such as True Positive and False Positive. You can now see triage outcomes directly in these views without navigating to individual cases, improving the workflow efficiency for alert analysis.

Detection/ML

Improvements

  • Improved External/Internal Suspected SQL Injection alert types by adding a Key Field listing suspected queries from the source host observed in the time window. The top 5 are shown when more than 5 are detected.

  • AELDEV-68304: Updated the User Login Location Anomaly to prioritize login records that include a username.

    Updated the User Login Location Anomaly to prefer login records that include a username when multiple records are available for the same login event. Previously, the detection might use a record containing only a system identifier such as a Windows SID, causing alerts to display an unreadable identifier instead of a username. User Login Location Anomaly alerts now display a recognizable username when one is available in the correlated login data. Existing prioritization of Entra ID and MFA records continues to take precedence over this change.

  • AELDEV-68301: Added ASN enrichment fields to Impossible Travel Anomaly.

    Added srcip2_asn and srcip2_asn_org fields to the Impossible Travel Anomaly, exposing ASN information for the second source IP address. This update lets you use ASN-based filtering, automation, and alert exclusion logic with Impossible Travel detections.

  • Added 6 alert integrations through Microsoft's Graph Security API. The alert integrations are: Microsoft App Governance, Microsoft Data Loss Prevention, Microsoft Defender for Identity, Microsoft Defender for Office 365, Microsoft Entra ID Protection, and Microsoft Threat Intelligence. See Integration of Third Party Native Alerts.

  • AELDEV-61011: Improved user counting accuracy for Microsoft Entra ID environments.

    Improved user counting accuracy for Microsoft Entra ID environments by reducing duplicate counts when the same user has multiple email addresses. This update helps make license-related user counts more accurate.

    This feature is available as part of the Early Access Program (EAP) for customers who requested it.

  • AELDEV-59125: Added alert suppression options for Impossible Travel and User Login Location anomalies.

    Added alert suppression options for the Impossible Travel Anomaly and User Login Location Anomaly when customizing these alert types in the Detection Management System. Select Attributes for DMS Alert Suppression to set how profile-based alert suppression in the Detection Management System (DMS) interacts with alerts from the Impossible Travel and User Login Location anomalies. Previously, alert suppression applied a single scope across all detections, one alert per user per interval, which caused the global alert suppression to interact with each detection's own alert suppression in undefined ways. For the Impossible Travel Anomaly, if Alert Suppression is enabled in DMS with a defined interval (default: 24 hours), the following attributes will rate limit or filter the alerts from this detection: Unique User and Unique Locations For User. For Unique User, a maximum of one alert per user will be triggered within the interval set in DMS. This is the previous behavior. For Unique Locations For User, a maximum of one alert per unique pair of locations per user will be triggered within the interval set in DMS. There may be multiple alerts for a single user. The User Login Location Anomaly also has Unique User as well as Unique Location For User, in which a maximum of one alert per location per user will be triggered within the interval set in DMS. See Using the Detection Management System.

Stellar Cyber Platform

New Features

  • Implemented detailed reason codes for condition evaluation when you use array comparisons. New reasons include ARRAY_COMPARE_PATH_MISSING, ARRAY_COMPARE_PATH_EMPTY, ARRAY_COMPARE_SUBPATH_MISSING, ARRAY_COMPARE_QUANTIFIER_NOT_MET, and ARRAY_COMPARE_MATCH_FOUND. Each reason provides a precise message indicating whether the condition path was not found, exists without buckets, has no matching values, or matches with a given quantifier. These reasons help you diagnose mismatches in aggregation query paths and verify that conditions are functioning correctly.

  • Enabled OpenAPI 3.1 output for the Auto Triage API. The generated swagger.json conforms to JSON Schema draft 2020-12 and supports $ref to external schemas. Existing paths and operations remain unchanged. Access the specification at /swagger.json for integration with code generators and validation tools. Regenerate software development kit (SDK) artifacts or update validation tooling to adopt OpenAPI 3.1 features.

  • Added OpenAPI 3.1 document generation with JavaScript Object Notation (JSON) Schema draft 2020-12 compatibility and external $ref support. Models, parameters, and responses now include expanded type information for more precise validation and client and server generation. Tools that support only OpenAPI 3.0 must be updated before consuming the document. You can access the specification at /swagger.json or through the API documentation page.

  • AELDEV-62965: Implemented API actions for license enforcement and usage notifications.

    Introduced API capabilities to create actions for license enforcement and usage monitoring, including notifications for exceeding asset and ingestion thresholds. This helps automate responses and manage licensing effectively through remote APIs.

Improvements

Sensors

New Features

  • Added support for installing the Linux Server Sensor on Fedora Linux 40 and later. Previously, the installer did not apply the required compatibility libraries on Fedora 40, causing sensor services to fail to start after installation. The installer now correctly handles Fedora 40 and later releases using the same compatibility library approach used for other supported Red Hat-based platforms.

  • Added support for installing the Linux Server Sensor on Oracle Linux 10. Systems running Oracle Linux 10 can now connect to Stellar Cyber and begin collecting data without requiring additional configuration changes.

Improvements

  • Updated the Deep Packet Inspection (DPI) protocol bundle to version 1.830.0-21. This update adds support for extracting prompts from decrypted traffic associated with AI tools such as ChatGPT, Copilot, and Claude, enabling more detailed traffic categorization and analysis for these platforms. The DPI engine version remains at 5.10.0-47.

  • Upgraded Suricata to version 8.0.1, which resolved vulnerabilities related to the handling of entropy, ldap.responses.attribute_type, and tls.subjectaltname keywords, and detection bypass issues. These updates enhance stability and security by rectifying segmentation fault and buffer overflow occurrences in earlier versions.

  • Secured internal socket communications used by sensor services on Linux Server Sensor (agent) deployments. Previously, these communications used unauthenticated connections, which could allow non-privileged users on the host server to perform unauthorized actions such as restarting services or triggering upgrades. Local service communications now require authentication, ensuring that CLI commands and service operations are accessible only to authorized users.

Connectors

New Features

  • AELDEV-61583: Introduced the Check Point Smart-1 Cloud connector.

    Added the Check Point Smart-1 Cloud firewall connector. This responder supports blocking/unblocking IP addresses across gateways managed by Check Point Smart-1 Cloud by manipulating host objects and groups referenced in firewall access policies. See Configuring Check Point Smart-1 Cloud Connectors.

  • AELDEV-61569: Introduced the Liongard connector.

    Added the Liongard connector to ingest devices, detections, and identities data though the Liongard API. See Configuring Liongard Connectors.

  • AELDEV-46279: Introduced the Ironscales connector.

    Added the Ironscales connector to query data related to incidents through the Ironscales Management API. See Configuring Ironscales Connectors.

Improvements

Parsers

New Features

  • Added a built-in parser for ingesting Radware WAF logs in JSON format on port 6208. The parser extracts web application firewall event fields including source IP addresses, destination hosts, attack names, threat categories, rule identifiers, and application details. This parser improves visibility into web application attack activity and supports investigation of threat events and policy enforcement.

  • Added a built-in parser for ingesting ManageEngine Endpoint Central logs in RFC 5424 syslog format on port 6203. This parser extracts endpoint management event data including endpoint identifiers, event categories, and activity details. This parser improves visibility into endpoint management activity and supports investigation of endpoint configuration and compliance events.

  • Added a built-in parser for ingesting Safetica Data Loss Prevention (DLP) logs in RFC 5424 syslog format on port 6202. This parser extracts DLP event fields including event ID, type, user, computer, rule name, and event details. This parser improves visibility into data loss prevention activity and supports investigation of policy violations and user behavior related to sensitive data.

  • Added a built-in parser for ingesting ONV Switch logs in RFC 5424 syslog format on port 6100. This parser extracts switch event data including source and destination IP addresses, ports, and protocol information. This parser improves visibility into ONV Switch network activity and supports traffic analysis and investigation.

  • Added a built-in parser for ingesting Oracle Audit Trail logs in JSON format on port 6204. This parser extracts database audit fields including user identity, database actions, and affected objects. This parser improves visibility into Oracle database activity and supports investigation of privileged access, schema changes, and data access events.

  • Added a built-in parser for ingesting Forcepoint Web Security (Filebeat) logs in RFC 5424 syslog format on ports 6101–6201. This parser extracts web security event data including source IP addresses, ports, and web security category information. This parser improves visibility into web traffic filtering activity and supports investigation of policy enforcement and blocked request events.

  • Added a built-in parser for ingesting Trend Micro Worry-Free Business Security Service logs in CEF format with a custom header on port 6209. The parser extracts endpoint security event fields including threat detections, source and destination IP addresses, ports, protocols, rule names, actions, and device identifiers. This parser improves visibility into endpoint security activity and supports investigation of threat events and policy enforcement across managed endpoints.

  • Added a built-in parser for ingesting Avaya IP Office logs in RFC 3164 syslog format on port 6099. This parser extracts VoIP event data including source IP addresses, ports, and call activity information. This parser improves visibility into Avaya IP Office telephony activity and supports traffic analysis and investigation of VoIP events.

  • Added a built-in parser for ingesting Custom Application JumpTo logs in JSON format on port 6098. This parser extracts application event data including source IP addresses, ports, and session activity. This parser improves visibility into JumpTo application activity and supports traffic analysis and investigation.

  • Added a built-in parser for ingesting Cowrie SSH/Telnet Honeypot logs in JSON format on port 6206. The parser extracts honeypot event fields including attacker source IP addresses, event types, session identifiers, captured credentials, and commands entered during sessions. This parser improves visibility into honeypot activity and supports investigation of unauthorized access attempts and attacker behavior.

  • Added a built-in parser for ingesting Kaspersky Secure Mail Gateway logs on port 6205. This parser extracts email security event data including sender and recipient information, threat classifications, and policy actions. This parser improves visibility into email security activity and supports investigation of malware detections, spam filtering, and blocked message events.

  • Added a built-in parser for ingesting Fudo Privileged Access Management (PAM) logs in RFC 3164 syslog format on port 6096. This parser extracts PAM event data including user identity, session activity, and access details for privileged account operations. This parser improves visibility into privileged access activity and supports investigation of administrative sessions and access policy enforcement.

  • Added a built-in parser for ingesting Ubika Web Application Firewall (WAF) logs in RFC 5424 syslog format on port 6097. This parser extracts WAF event data including source IP addresses, ports, HTTP methods, and web security category information. This parser improves visibility into web application firewall activity and supports investigation of attack attempts, policy violations, and blocked requests.

Improvements

  • Updated the Fortinet FortiWeb built-in parser to parse the http_request_bytes and http_response_bytes fields as numeric types rather than strings. The parser previously emitted these fields as string values, which prevented their use in ATH calculations. They are now available as numeric fields, letting you use HTTP request and response byte counts in threshold-based detections and analytics.

  • Updated the Microsoft Internet Information Services (IIS) built-in parser to handle log entries that include an extended field set, in addition to the field layout supported in earlier releases. The parser extracts source and destination IP addresses, ports, HTTP method, URL, and response code from both field layouts. The extended field set additionally includes the URI query string, authenticated username, HTTP referrer, and request and response byte counts, providing greater visibility into web request patterns and user activity. These additional fields are now normalized and searchable, supporting more detailed investigation of IIS web traffic.

  • DATA-3463: Improved Citrix ADC login failure event parsing.

    Improved Citrix ADC parser support for login failure events by extracting additional details from the event description, including the client IP address and username. The update also added parsing for browser type when present in the log. This enhancement provides better normalization for Citrix ADC authentication failures and improves investigation context for user activity and source attribution.

  • Expanded the A10 vThunder Application Delivery Controller (ADC) parser to support additional firewall syslog traffic log formats, including FW-TCP-*, FW-UDP-*, and FW-IC4-* events that use src_ip:src_port<-->dst_ip:dst_port message patterns and embedded key-value fields such as ACT (action), RT (timestamp), IN-INTF, OUT-INTF, POLICY, RULE, and DUR (duration). This update improves parsing and normalization for A10 traffic and session events so they can be investigated more consistently.

  • Updated the Cisco ACI parser to handle additional syslog header patterns for fault, event, session, and NX-OS–style messages. The update also added support for more Cisco ACI SYSTEM_MSG event types and improved extraction of details from events such as VMware controller connection failures, SOAP login errors, job status changes, NTP faults, and fabric link changes. This improvement helps ensure that more Cisco ACI logs are parsed and normalized correctly for investigation.

  • Updated the Arbor Peakflow SP parser to handle additional syslog header variants, including RFC 3164, RFC 5424, and Arbor custom header formats. The update also added support for more Arbor event message patterns and improved extraction of details from alert lifecycle, router leadership, interface threshold, flow-count, and configuration change events. This improvement helps ensure that more Arbor Peakflow SP logs are parsed and normalized correctly for investigation.

  • Extended support to include Snort-formatted intrusion event logs in the KVH CommBox Edge Gateway parser. The update parses Snort alert identifiers, message text, Classification, Priority, protocol, and source and destination IP address and port information, including repeated-message wrappers. This improvement helps ensure that Snort detection events forwarded by KVH CommBox Edge Gateway devices are parsed and normalized correctly for investigation.

  • Moved additional Fortinet FortiAnalyzer parser fields from msg_data into the Fortinet vendor field namespace. This update improves field organization and makes those Fortinet-specific values easier to access consistently during search, analysis, and downstream content development.

    advpnsc dstserver init rcode
    bandwidth dstswversion lease role
    cloudaction error masterdstmac setuprate
    clouduser espauth mem srcfamily
    cpu esptransform mode srchwversion
    dhcp_msg exch nas stage
    dir faclogindex newvalue sysuptime
    disk fazlograte oldvalue totalsession
    disklograte freediskstorage out_spi useralt
    dstdevtype healthcheck peer_notif userip
    dsthwvendor in_spi probeproto waninfo
    dstosname      

  • Improved the Common Event Format (CEF) parser for FortiWeb logs by promoting recognized FortiWeb extension fields to the Fortinet vendor namespace instead of leaving them in generic message data. This update makes important web application firewall context, such as request details, client application data, device identifiers, outcomes, and reasons, easier to access consistently in searches, detections, and investigations.

  • Added support for RFC 3164-formatted syslog messages in the Penta Security WAPPLES parser, including key-value log content sent by the web application firewall. This update improves compatibility with WAPPLES syslog output and helps ensure that more web application firewall and system monitoring events are parsed consistently.

  • Moved the bssid field from the top level into the Fortinet vendor namespace in the Fortinet FortiGate parser. This update improves field organization by keeping FortiGate-specific wireless access point information with related vendor fields, which makes the data easier to interpret and use consistently in searches, content, and investigations.

  • Updated the VMware ESXi built-in parser to normalize additional fields from authentication-related log events, including user name, source IP address, login result, and event action. These fields were previously unparsed and unavailable for search or investigation. They are now normalized and searchable, improving visibility into ESXi authentication activity and supporting investigation of login events and access patterns.

  • Added support for BSD syslog-style headers in the Citrix ADC parser, including logs that use DD/MM/YYYY:HH:MM:SS GMT timestamps instead of RFC 5424 headers. This update improves compatibility with Citrix ADC SSLVPN session and connection statistics logs, such as TCPCONNSTAT and UDPFLOWSTAT, so those events can be parsed more reliably for access monitoring and troubleshooting.

  • Updated the Efficient IP parser with enhancements to improve performance and prevent potential BufferOverflow conditions. These changes include time parser optimizations for GC efficiency, expanded regex for client IP matching to capture non-standard IP formats, and improved record normalization processes. Support for correct base class initialization logic was added to align parser behavior with the implemented processing paths. These updates address performance and normalization requirements, effectively reducing risk without compromising processing integrity.

  • Added support for a new HTTP-ingested JSON log format in the HTTP Google Kubernetes Engine parser, including Kubernetes container logs forwarded with nested docker and kubernetes metadata. This update improves compatibility with Google Kubernetes Engine log payloads so container, pod, namespace, host, and image context can be ingested more reliably for Kubernetes monitoring and investigation.

  • Added support for an additional Arbor Peakflow SP log format used for blocked-host threat events generated by arbor-networks-aps. The update adds parsing for ATLAS threat-category logs and extracts related threat context, including threat name, threat category, protection group details, match type, and reference identifiers. This improvement helps surface Arbor blocking and reputation activity more clearly.

  • Added a "This field is required" validation message for the Custom Namespace field in Parser Studio when it is left empty.

  • Added Raw Log Capture support for modular parsers in Parser Studio, aligning modular parser capabilities with the existing feature set available for legacy parsers.

  • Removed the HTTP JSON parser from the Parser Studio parser list. The HTTP JSON parser continues to use a sensor profile configuration and is not managed through Parser Studio.

  • AELDEV-71104: Set new built-in parsers added in 6.5.0 and later to inactive by default.

    Built-in legacy parsers added in 6.5.0 and later releases are now inactive by default for new tenants and must be explicitly activated in Parser Studio before use. Previously added legacy parsers retain their existing active status and are not affected by this change.

  • Updated the show logforwarder CLI command to show readable names for modular parser activity counters. When you view parser activity in the command output, input counts now display the parser name and output counts now display the index name instead of internal identifiers. This update makes it easier to understand which modular parsers are receiving data and where that data is being sent.

  • AELDEV-68064: Updated the Parser Studio UI to support enabling parser ports on demand.

    Updated the Parser Studio user interface to support enabling and disabling parser ports on demand. This capability is available to all users, while the broader Early Access Program-only Parser Studio features for creating and managing custom parsers remain limited to customers participating in the Early Access Program.

  • Added statistics to the show logforwarder command for modular parsers, including input record counts aggregated by parser and output record counts aggregated by index. This update extends CLI visibility to modular parser pipelines and reports counters for input and output activity. It helps you validate log flow more quickly, confirm parser health, and identify configuration issues or dropped-record conditions.

  • Enhanced the Normalization step in Parser Studio to support the automatic detection and normalization of multiple timestamp formats, including epoch in seconds and milliseconds, ISO 8601, and other common date and time string formats. A manual override option was added that lets you specify the timestamp field and format when automatic detection does not produce the expected result. This update improves timestamp handling accuracy and helps ensure that ingested events appear at the correct time in the Stellar Cyber Platform.

  • AELDEV-56360: Enabled selective parser port activation to reduce false alerts.

    Enabled selective activation and opening of parser ports per tenant.You can now choose which parser ports are active in Parser Studio, which helps prevent unwanted data from being ingested and reduces false alerts caused by unused listeners.

Usability

New Features

  • Resolved an issue in which manually triggered rules were incorrectly shown as inactive failures in their status display. Manual trigger statuses are now recorded accurately and reflect the actual input status.

  • AELDEV-68524: Added new platform health alerts to the System Action Center.

    Integrated platform monitoring alerts into the System Action Center to improve early detection and response coordination. Alert types include ingestion drops, authentication errors, and cluster health issues. Alerts include severity values, support filtering by time range, severity, and type, and can be enabled or disabled individually per tenant. Alert records are retained for at least 30 days.

    This feature is available as part of the Early Access Program (EAP).

  • AELDEV-65225: Enabled Add to Watchlist for all alert fields.

    Removed the field restriction and enabled Add to Watchlist for all alert fields. You can now select any value in Alerts | Alert Detail and add it to a watchlist. Supported types include IP, Domain, URL, Hash, Hostname, and Username. Fields such as host.ip and dns.question.name now qualify automatically. Entries that do not match the selected type display a validation message.

  • Added the option to attach a JSON-formatted file containing filtered calculation results to alert notification emails that ATH calculation rules generate when triggered. The JSON attachment contains the same filtered records as the CSV attachment, limited to entries that meet the calculation condition threshold. Including the results as an attachment in JSON format lets you process or integrate ATH calculation results programmatically in addition to reviewing them in spreadsheet format.

  • AELDEV-57363: Updated auto-triaged alert details with a redesigned view.

    Updated the auto-triaged alert details view with a redesigned layout that preserves existing alert information and investigation workflows. The updated view continues to provide access to key fields, details, JSON view, rules, and comments and activity, while adding new Auto Triage sections such as a Simple Summary and Findings to make results easier to review.

  • AELDEV-53674: Added a response action panel on the Auto Triage alert page.

    Added a response action panel to the Auto Triage alert page. You can initiate supported response actions directly from an alert without leaving the triage view. Available actions match the response actions configured through your current integrations. Action execution records appear on the Activity page for audit and follow-up.

Improvements

Early Access Program

If you're interested in testing out new features ahead of general availability, consider joining the Early Access Program (EAP) by contacting your Stellar Cyber Customer Success representative and telling them which EAP feature you want to test. Once you've agreed to the EAP terms and signed up, the EAP feature is unlocked for you.

The purpose of this program is to boost performance and reliability through real-world customer insights, giving you a hands-on role in shaping a Stellar Cyber feature. In return, you'll receive early access to upcoming releases and the chance to guide product development.

The following EAP features are in this release:

AI Assistant

AI Assistant is a natural-language investigation workbench built into every automatically triaged case. It converts your plain-language questions into answers grounded in the specific case data, alert details, and Verdict Signal Check (VSC) context that Auto Triage has already gathered. While the automated verdict tells you what Stellar Cyber concluded, AI Assistant explains why, and it supports follow-up questions in your own words, with no query syntax to learn. AI Assistant is included with every Auto Triage license at no additional cost and is available on SaaS deployments only.

Exportable Dashboards — Report Integration

Exportable Dashboards lets you schedule dashboards created with the Dashboard Builder as recurring PDF reports. The schedule form includes a new Default PDF type that renders the dashboard as it appears in the Dashboard Builder, along with options to control table row counts, chart color palettes, and optional CSV exports. This capability lets you generate consistent, configurable reports from new dashboards while preserving the settings and PDF types used by existing scheduled reports.

MCP Server

The Stellar Cyber MCP Server connects supported AI clients to the Stellar Cyber Platform through the Model Context Protocol (MCP). The MCP server lets AI clients retrieve case and alert data, review investigation context, perform tenant-aware operations, and update selected case fields. This capability helps teams extend AI-assisted investigations by giving approved clients structured access to operational security data and workflows.

Parser Studio

Parser Studio lets you create and manage custom log parsers for data ingestion by cloning existing parsers, testing parser behavior before deployment, and activating parsers for production use. This capability helps you accelerate onboarding of custom log sources while reducing parser development effort and improving validation before live ingestion.

XDR Connector Webhook Ingestion

This is a simple webhook framework that lets you post JSON data directly from any external system into Stellar Cyber, accelerating custom integrations and expanding your visibility across the entire security stack. The XDR Connector is in Public Preview in this release.

Customizable Case Correlation Strategies

This EAP feature introduces support for multiple case correlation strategies, allowing teams to evaluate and experiment with different approaches to grouping alerts into cases. Each strategy provides a distinct investigative perspective:

  • Attacker-Centric Correlation groups alerts by the source (attacker) host, making it easier to track adversary behavior across multiple targets.

  • Victim-Centric Correlation organizes alerts by the destination (victim) host, enabling focused protection and visibility on high-value assets.

  • Multi-Entity Correlation links alerts across interconnected hosts and actions to form a single case, offering a holistic view of extended or lateral attack campaigns.

This flexibility enables security teams to tailor investigations based on their operational priorities—whether that’s identifying persistently targeted endpoints, tracing threat actor movements, or capturing full-scale intrusion campaigns.

Alert for Suspicious OCI Tenant-to-Tenant Communication

This EAP feature introduces a new alert type that detects cross-tenancy communications in the Oracle Cloud Infrastructure (OCI). By analyzing tenantId fields in audit logs, the feature identifies requests that target resources in a different tenancy. This provides accurate visibility into potentially unauthorized cross-tenancy activity and strengthens oversight in OCI environments.

To join the Early Access Program and begin testing these features, contact your Stellar Cyber Customer Success representative.

Resolved Issues

The following issues have been resolved in this release.

  • Fixed a normalization issue in the CEF parser that prevented the user_agent field in Check Point logs sent to port 5143 from being retained in parsed data. The issue occurred because the previous normalization logic did not correctly handle mapping user_agent to user_agent.original. After this fix, logs that include user_agent preserve the value correctly so it remains available for investigation and search.

  • Resolved a parsing issue in the Netscaler parser that incorrectly removed double quotation marks contained within the command field value. The parser now preserves quotation marks that are part of the original command content while still handling wrapper quotation marks correctly. This fix ensures that parsed command values remain accurate for investigation and review.

  • Resolved three security vulnerabilities in the restricted CLI shell on Data Processor deployments. An authenticated CLI user could reach unprotected code paths that allowed operating system commands to run with elevated privileges. A local service interface was also reachable without authentication, providing an additional path to execute arbitrary commands. All three attack paths have been closed, and the restricted CLI shell now enforces access controls consistently across all commands and interfaces.

  • Resolved a task status tracking issue in which uninstall tasks for Windows Server Sensors remained indefinitely in an In Progress state even after the sensor had successfully uninstalled. The sensor was not sending a required identifier in its completion message, so the Central Manager could not match the response to the original task and never updated its status. Uninstall tasks now complete and reflect the correct status after the sensor is removed.

  • Resolved two related issues that occurred after you uninstalled a Linux Server Sensor from the Stellar Cyber Platform UI. The sensor software directory was not fully removed, leaving files on disk after the uninstall completed. Additionally, the system package database continued to show the sensor package as installed rather than reflecting its removed state. Both issues have been corrected: the sensor directory is now fully removed on uninstall, and the package status correctly reflects the uninstalled state.

  • Resolved an upgrade issue on Modular Sensors running Ubuntu 22.04 in which the upgrade process stopped responding and never completed. The upgrade stalled while attempting to remove a legacy package repository reference that required contacting an external server. If the sensor had limited or no access to that server, the process waited indefinitely, leaving most sensor services down and the upgrade showing no result. The upgrade process no longer waits indefinitely on this step and completes correctly.

  • Resolved a session management issue on on-premises Data Processor deployments. When a timed support session expired while you had an elevated root shell open, the session did not close cleanly. Instead, the CLI and the root shell interleaved, producing repeated error output on every keystroke and preventing you from exiting normally. Support sessions now terminate cleanly regardless of whether an elevated shell is active when the session expires.

  • Resolved CVE-2026-41651, a security vulnerability in the PackageKit component present on sensors running Ubuntu 22.04. The installed version of PackageKit fell within the affected range, and the patched version was not available through the previously configured update source. The sensor package has been updated to include the patched version, and sensors can now apply the fix using the exec cve patch CVE-2026-41651 CLI command. Vulnerability scanners such as Tenable and Qualys no longer flag this CVE on updated sensors.

  • Resolved an issue in which the file-inspection pipeline on Modular Sensor deployments stopped capturing and analyzing files. The issue occurred because a build configuration error caused Suricata to use an inefficient pattern-matching method that consumed excessive memory, eventually causing the capture process to stop. Newly built sensors now include the correct configuration, allowing file inspection to operate reliably within expected memory limits.

  • Resolved an inconsistency in which the show logforwarder external-server command was available on Network Sensors and Security Sensors but not on Modular Sensors. If you configured an external syslog forwarding server on a Modular Sensor, you had no way to verify the configuration from the CLI. The command is now available on Modular Sensors and returns the configured external server address as expected.

  • Resolved a display issue in which InSync configurations you had deleted continued to appear on the Alerts page even after they had been successfully removed. The Alerts page was retrieving InSync status data from a source that was not updated on deletion, so removed entries continued to show with a Paused status. Deleted InSync configurations no longer appear on the Alerts page after you remove them.

  • Resolved an issue in which the file-inspection pipeline produced incorrect results when you transferred two files over a Network File System (NFS) share in rapid succession using NFSv3. The second file was not fully captured, resulting in an incomplete scan and an incorrect file hash. NFSv4 transfers were not affected because that version handles file boundaries differently. The fix ensures each file is inspected completely and independently regardless of how quickly transfers follow one another.

  • Resolved a memory management issue on sensor deployments where available RAM was limited, particularly on 8 GB hosts. Under sustained load, memory consumption across sensor services exceeded available memory, causing traffic inspection to crash and restart repeatedly. Sensors on VMware ESXi hosts were additionally affected when the hypervisor reclaimed guest memory under host-level pressure. Memory limits are now calculated dynamically based on the available RAM of the host, preventing the crash cycle and keeping traffic inspection running reliably.

  • Resolved a bulk case action issue in which a typed tag was not applied if you selected Apply without first confirming the entry. The tag input is now committed before the bulk action request is sent, so the complete tag list is applied to the selected cases. Existing tag workflows that use Enter or an autocomplete selection continue to work as before.

  • Resolved multiple performance issues on Modular Sensor deployments that caused traffic inspection to degrade or fail under sustained high traffic. Under heavy load, processing threads competed for resources inefficiently, memory usage grew in a cycle that repeatedly disrupted normal operation, and sessions were exported without full inspection data. These issues combined to reduce the amount of traffic the sensor processed reliably. Processing thread allocation, memory management, and session handling have all been improved, resulting in more stable operation and significantly higher throughput under high traffic conditions.

  • Resolved an issue in which the Cases REST API endpoint stopped responding when a request included one or more cases in the result set. Requests for case counts and metadata continued to work normally, but any request that required retrieving actual case records would never return. This affected users and integrations that relied on the Cases API to retrieve case data, including SOAR integrations and automated workflows. The API now returns case data reliably for these requests.

  • Resolved an issue in which indices created for new tenants during periods of high onboarding activity did not receive the correct storage placement settings. When data arrived for a new tenant before the regular index creation process completed, the resulting indices were placed without the required constraints, meaning the data could land on unsuitable nodes. Affected indices now receive the correct placement settings, ensuring data is stored on the appropriate nodes as expected.

  • Resolved an issue in which the set malware_upload, show malware_upload, and unset malware_upload CLI commands were visible and available on Network Sensors and Linux Server Sensors running in agent mode, even though those sensor types do not support malware detection. Running these commands on unsupported sensor types had no effect and could cause confusion. These commands are now shown only on sensor types that support malware detection.

  • Implemented automatic reload of log filters in the connector for Azure Event Hub. New or modified filters take effect during active ingestion without a process restart and no longer raise exceptions.

  • Resolved an issue in which installing a Linux Server Sensor on Amazon Linux 2023 displayed a "No such file or directory" error during the setup process, even when the installation completed successfully. The message appeared because the installer attempted to write to a scheduler configuration file that did not exist on that platform. The installer no longer attempts this step on Amazon Linux 2023, and the installation now completes without the misleading error.

Known Issues

Upgrading Sensors

Depending on the type of server sensor, you can upgrade your sensors directly to version 6.6.0 from these previous versions:

  • Linux Server Sensors: 6.4.0 or 6.5.0

  • Windows Server Sensors: 5.1.0 through 6.5.0

Upgrade the sensors to version 6.6.0 using the following process:

  1. Prepare for the upgrade.

  2. Upgrade the sensors.

  3. Verify the upgrade.

Prepare for the Upgrade

To prepare for the upgrade:

  • Make sure the sensors are up and running
  • Take note of the ingestion rate for the sensors to be upgraded in the Sensor Details page
  • Make sure the system health indicators in the Sensor Details page all show green.

Upgrade the Sensors

New features, updated ML algorithms, and enhanced configurations may change ingestion and detection patterns. We recommend the following to ensure a smooth upgrade:

  • Upgrade sensors with the Sandbox and IDS features enabled before sensors with the only the Network Traffic feature enabled. Sensors with Network Traffic enabled send data to sensors with Sandbox and IDS enabled for additional processing.
  • Upgrade sensors in batches instead of all at once.
  • For server sensors (agents):
    • Upgrade a small set of sensors that cover non-critical assets.
    • After 24 hours, ensure that your ingestion is as expected, then upgrade a larger set.
    • After 24 hours, ensure that your ingestion is as expected, then upgrade the remaining server sensors.

CentOS 7.1 Prerequisite – Update curl to 7.29.0-59.el7_9.2 or Higher

Before upgrading any Linux Server Sensors running in CentOS 7.1, you must check your curl version and update it to 7.29.0-59.el7_9.2 or higher to use the strong encryption required by the Stellar Cyber Platform.

  1. Check your curl version as shown below:

    yum list installed curl

    \* Loaded plugins: fastestmirror Loading mirror speeds from cached hostfile Installed Packages curl.x86_64 7.29.0-19.el7

  2. If the listed version is lower than 7.29.0-59.el7_9.2 (as it is in the example above), use the following commands to update the curl package:

    yum makecache

    yum install curl

  3. If installation of the curl package fails, it is most likely because CentOS is trying to use a repo that has reached its end of life. Try updating the base URL and then reinstall curl. The following sed command makes the necessary changes for most environments to ensure that the updated curl package can be installed:

    sudo sed -i.bak -e 's|^mirrorlist=|#mirrorlist=|' -e 's|^#baseurl=http://mirror.centos.org/centos/\$releasever|baseurl=http://archive.kernel.org/centos-vault/7.9.2009|' /etc/yum.repos.d/CentOS-Base.repo

To upgrade sensors:

You can upgrade a sensor to the most recent release from the two previous releases. This means that you can upgrade a sensor to the 6.6.0 release from any 6.4.x or 6.5.x release.

If you are upgrading a Windows Server Sensor, complete any pending updates for the host Windows machine before upgrading the Server Sensor.

  1. Select System | DATA SOURCE MANAGEMENT | Sensors | Sensors.

    The Sensor List appears.

  2. Select Manage | Software Upgrade.

    The Sensor Software Upgrade page appears.

  3. Choose the target software version.

  4. Choose the target sensors.

  5. Select Submit.

Verify the Upgrade

To verify that the upgrade was successful:

  • Check the Software Version in the Sensor List.
  • Check the Sensor Status LED in the Sensor List.
  • Check the ingestion rate in the Sensor Details page for upgraded sensors and make sure it is as expected.