Skip to main content
Glama
KasperskyLab

Kaspersky OpenTIP MCP Server

Official
by KasperskyLab
CHANGELOG.md4.78 kB
# What's new in Kaspersky TI connector for OpenCTI ## [1.1.0] - 2025-12-06 ### Upgrade notes This release changes how indicators, observables, relationships and TAXII incremental collection windows are handled. To make sure your OpenCTI graph is consistent and no data is lost, you **must** reinitialize the connector after upgrading: - stop the existing connector instance - reset the connector state in OpenCTI (Admin → Data → Connectors → your Kaspersky connector → Reset state) - start the connector again so it performs a fresh synchronization with the updated logic --- ### Indicator names - Indicator names are now derived directly from the STIX pattern value. - For `URL LIKE` patterns, SQL-style wildcards `%` and `_` are converted to `*` and `?` to produce more readable, mask-like indicator names. ### Observable generation - The connector now explicitly generates observables from indicator patterns. - Supported observable types: `url`, `domain-name`, `ipv4-addr` and `file`. - Fixed observable generation for `file` objects that contain multiple hashes. - Added support for patterns with `LIKE`, including correct handling of wildcard masks. - All indicators are linked to the corresponding observables via the `based-on` relationship. ### Confidence handling - A `connector.confidence_level` configuration parameter controls the `confidence` value for entities and relationships created by the connector. - The configured confidence value is applied to all relationships created by the connector (for example, between indicators, countries and threat actors). - Confidence values that come directly from the feed (for indicators or IPs) are not overridden. ### Richer relationships in the graph - Relationships created by the connector (including `indicates` and `based-on`) now inherit part of the metadata from the source objects: - `description` - `labels` - This makes the OpenCTI graph easier to analyze and filter, since relationship metadata can now be used in searches and views in the same way as object metadata. ### Threat score and tagging - A single `connector.threat_score` parameter now controls the `x_opencti_score` assigned to all created indicators and their related observables (URLs, IPs, hashes, etc.). If you configure the score as `80`, all newly created indicators will receive `x_opencti_score = 80`, regardless of type. - The `connector.threat_score_high` and `connector.threat_score_medium` thresholds drive automatic threat-level tags: - `threat_score:kaspersky:high` - `threat_score:kaspersky:medium` - `threat_score:kaspersky:low` - If an object description contains a fragment like `threat_score=NN`, the connector parses this value and assigns the corresponding threat-level tag based on the configured thresholds. ### Activity label normalization - Activity labels from the feeds are normalized for consistency: - `malicious-activity` is automatically converted to `malicious-activity:kaspersky`. - This simplifies filtering by activity type and source in OpenCTI. ### TAXII ingestion and reliability - A new `kaspersky.connection_timeout` configuration parameter defines the HTTP timeout (in seconds) for all TAXII requests. - Object retrieval from TAXII collections now uses automatic retries: - up to 3 attempts on transient network errors - with logging and increasing delay between attempts - The incremental collection window has been reworked to avoid data loss when TAXII delivers objects "in the past": - previously, the connector fetched “last hour” of data every hour, which could miss indicators with backdated timestamps - now, the connector requests new data starting from the `created` timestamp of the most recently ingested indicator, ensuring late-arriving indicators are still collected ### OpenCTI / pycti compatibility - Threat actor ID generation is adapted for pycti 6.4+ while preserving the previous behavior for older versions. - This prevents duplicate threat actors and ID conflicts across OpenCTI versions and keeps the connector compatible with both new and legacy deployments. ### New configuration parameters The following parameters are available in `config.yml` and via environment variables: - `connector.confidence_level` Controls the `confidence` value for entities and relationships created by the connector. - `connector.threat_score` Controls the `x_opencti_score` assigned to all created indicators. - `connector.threat_score_high`, `connector.threat_score_medium` Define thresholds used to assign threat-level tags (`threat_score:kaspersky:high/medium/low`). - `kaspersky.connection_timeout` HTTP timeout for TAXII requests in seconds; used by all TAXII client operations, with automatic retries on failures.

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/KasperskyLab/threat-intelligence'

If you have feedback or need assistance with the MCP directory API, please join our Discord server