MCP Tool Reference¶
Use relic_mcp_tool_reference for the live tool list exposed by the running
server.
Result Limits¶
MCP tools intentionally return bounded result sets. A bounded response is not evidence of absence.
result_limitmeans a limit was active.result_limit_warningmeans the returned rows reached or appear to have hit a limit.- Increase the tool
limit, read an existing generated report/export, or ask for a dossier/full context before relying on absence.
Routing and Discovery¶
relic_route_questionrelic_mcp_workflow_guiderelic_mcp_tool_referencerelic_workspace_summaryrelic_workspace_maprelic_workspace_healthrelic_case_evidence_maprelic_case_readinessrelic_list_casesrelic_list_computersrelic_list_imagesrelic_list_jobsrelic_get_jobrelic_processing_progressrelic_resume_planrelic_profile_previewrelic_doctorrelic_discover_reportsrelic_discover_report_exportsrelic_read_existing_report
Case Review¶
relic_case_dashboardrelic_case_reviewrelic_case_activity_digestrelic_case_next_actionsrelic_case_runbookrelic_timelinerelic_timeline_windowrelic_activity_windows
Artifact Queries¶
relic_query_evidence_contentsrelic_query_filesystem_listingsrelic_search_artifactsrelic_search_contentrelic_get_indexed_contentrelic_lead_searchrelic_artifact_search_sourcesrelic_file_dossierrelic_usb_dossierrelic_user_activityrelic_query_suspicious_executionsrelic_query_external_storagerelic_query_usb_filesrelic_query_usb_contentsrelic_query_file_movement_identityrelic_query_opened_from_removable_mediarelic_query_opened_from_cloud_storagerelic_query_cloud_artifactsrelic_query_memory_artifactsrelic_query_browser_activityrelic_query_registry_activityrelic_query_shortcutsrelic_query_communicationsrelic_query_system_users
Packets and Jobs¶
relic_write_review_packetrelic_list_review_packetsrelic_read_review_packetrelic_write_search_packetrelic_list_search_packetsrelic_read_search_packetrelic_rerun_search_packetrelic_list_mcp_jobsrelic_get_mcp_jobrelic_get_mcp_job_outputrelic_get_mcp_job_progressrelic_list_progress_manifestsrelic_cancel_mcp_job
Gated Processing¶
These require --allow-processing:
relic_import_triage_ziprelic_import_report_bundlerelic_process_imagerelic_run_profilerelic_recover_deleted_files
relic_recover_deleted_files starts a tracked MCP job. The job record stores the
Perceptor CLI command, and the recovery output manifest stores the exact icat
command used for each recovered or failed candidate.
Reports¶
relic_list_report_typesrelic_generate_reportrelic_write_report_bundlerelic_report_bundle_coveragerelic_ingest_triage_zip_preflight
Use relic_generate_report with report_name: "bits-activity" for BITS,
qmgr, OneDrive updater, component updater, or transfer-job questions. The report
uses timestamped BITS Client EVTX rows and shows qmgr database/carved
correlations when exact job ID or URL matches are available.
Use relic_generate_report with report_name: "examiner-edge-artifacts" for
Sticky Notes, Windows notifications, NetworkList, outbound RDP history,
MountPoints2, Scheduled Task XML, CryptnetUrlCache, hosts, WSL, Windows Update,
Credential/Vault metadata, Bluetooth paired devices, installed applications, or
SwiftKey/InputPersonalization questions.
Use relic_generate_report with report_name: "mapped-network-paths" for
mapped network drives, UNC shares, or MountPoints2 keys that look like
##host#share#path. Perceptor decodes those keys into \\host\share\path and
returns the user profile plus first/last observed registry timestamps.
Use relic_generate_report with report_name: "non-standard-ads" for hidden
or alternate data stream questions. This report filters common
Zone.Identifier streams and classifies Cloud Files/OneDrive metadata, WOF
compression, SmartScreen, and NTFS metadata streams separately from
high-priority unclassified ADS rows.
Use relic_generate_report with report_name: "ntfs-security-descriptors" for
$Secure, $SDS, ACL, or NTFS permission-change questions. This report
inventories security descriptor streams from MFT ADS rows and clearly marks that
current output is presence/metadata-only, not decoded ACL content.
Use relic_generate_report with report_name: "remote-access-tool-logs" for
AnyDesk, TeamViewer, LogMeIn, ConnectWise Control, Splashtop, RustDesk, VNC, and
similar remote-support application log questions. Parsed log lines are
categorized into connection, authentication, transfer, and identity/routing
leads when possible.
Use relic_generate_report with report_name: "structured-memory" for
Volatility and MemProcFS structured memory questions. The report returns parsed
rows when available and also lists tool run attempts, failures, and no-row
results so analysts can distinguish "not run" from "run but unsupported for this
dump."
Use relic_generate_report with report_name: "event-interpretation" for
high-value EVTX questions involving account manipulation, log clearing,
PowerShell, scheduled tasks, WMI persistence indicators, print-service history,
service installs, or 4688 process creation.
Use relic_generate_report with report_name: "clipboard" for clipboard
history, copied/pasted content, or cloud clipboard sync questions. The report
uses the dedicated Windows clipboard store where available; Windows Activities
is secondary context.
MCP interactive query tools keep row limits for model usability. Direct
generated-report responses remain bounded, while saved report bundles default to
broader exports. If an MCP response contains result_limit_warning or a
generated report contains limited: true, read or regenerate the saved export
with a higher limit before making a negative finding.