skills/mukul975/anthropic-cybersecurity-skills/hunting-for-dcom-lateral-movement/Gen Agent Trust Hub
hunting-for-dcom-lateral-movement
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/agent.pyuses thesubprocess.runfunction to executereg querycommands. This functionality is part of theaudit_dcom_configfeature, which checks the local Windows registry for the registration of specific high-risk DCOM objects. The commands utilize static registry paths and a hardcoded list of class identifiers (CLSIDs). - [INDIRECT_PROMPT_INJECTION]: The scripts
scripts/agent.pyandscripts/detect_dcom_lateral_movement.pyingest and process Sysmon and Windows Security log files, which are external data sources potentially influenced by an adversary. - Ingestion points: Event log files are ingested through the
--evtxand--securitycommand-line arguments. - Boundary markers: The scripts utilize specialized libraries (
python-evtxandlxml) for structured XML parsing, but they do not use specific delimiters or instructions to isolate the parsed data when generating reports. - Capability inventory: The skill possesses capabilities for reading local log files, performing registry queries via subprocess, and writing JSON analysis reports to the file system.
- Sanitization: The scripts extract process metadata and command-line strings from logs for analysis and reporting but do not perform deep sanitization of these values before including them in the final output.
- [EXTERNAL_DOWNLOADS]: The skill documentation and script comments recommend the installation of the
python-evtxandlxmlpackages from standard software registries to enable log parsing capabilities.
Audit Metadata