CONDUIT TICKET REPORT

Finding Summary

Short description

[TITAN] High — databricks on dbw-banking-aml-notebooks

Severity

High

Priority

2 - High

Resource

dbw-banking-aml-notebooks

Resource type

Microsoft.Databricks/workspaces

Cloud

Azure

Subscription / Account

4f29d094-1079-44c9-acb0-4d73a7a2dd34

Resource group / Project

rg-banking-prod

The Security Finding

AML scoring Databricks workspace has 14 notebooks with hard-coded Snowflake passwords + 3 Unity Catalog tables exposing raw wire data with no row-level security. BSA/AML recordkeeping + FFIEC data-classification policy breach — insider-threat + regulator-visibility risk.

ITIL Change Management Fields

Justification

High: AML scoring Databricks workspace has 14 notebooks with hard-coded Snowflake passwords + 3 Unity Catalog tables exposing raw wire data with no row-level security. BSA/AML recordkeeping + FFIEC data-classification policy breach — insider-threat + regulator-visibility risk.

Implementation Plan

1. Pre-change snapshot captured by TITAN (auto-rollback available).
2. Execute fix command:
   databricks secrets put --scope aml-prod --key snowflake-pw && databricks unity-catalog grants update --full-name banking.aml_wires --row-filter amount_filter
3. TITAN FORGE verifies the fix was applied.
4. Post-change rescan by TITAN SCOUT — finding must no longer appear.
5. Close ticket with Successful close_code.

Risk & Impact Analysis

Risk level: MEDIUM-HIGH risk — misconfiguration with realistic exploit path.
Business impact if unremediated: Increases attack surface; auditor finding likely.
Scope: single resource (dbw-banking-aml-notebooks).
Blast radius: change is idempotent; pre-change snapshot captured by TITAN; auto-rollback available if rescan fails.
Finding detail: AML scoring Databricks workspace has 14 notebooks with hard-coded Snowflake passwords + 3 Unity Catalog tables exposing raw wire data with no row-level security. BSA/AML recordkeeping + FFIEC data-cla

Backout / Rollback Plan

1. TITAN auto-captured snapshot of dbw-banking-aml-notebooks before change (baseline: titan-banking-demo-20260422T201922Z).
2. If post-change rescan still shows the finding OR a new issue appears within 15 min:
   a. TITAN FORGE fires rollback automatically using stored snapshot.
   b. Incident reopens and escalates to on-call.
3. Manual rollback command path (human override) is documented in close notes.

Test Plan

1. TITAN SCOUT rescans dbw-banking-aml-notebooks immediately after FORGE applies the change.
2. PASS criteria: the specific finding no longer appears in SCOUT results.
3. PASS criteria: no new CRITICAL or HIGH findings introduced by the change.
4. Automated compliance check: HIPAA/PCI/SOC2 controls re-evaluated.
5. If any check fails, backout plan fires automatically.

Recommended Fix

databricks secrets put --scope aml-prod --key snowflake-pw && databricks unity-catalog grants update --full-name banking.aml_wires --row-filter amount_filter

Compliance Mapping

Routing Metadata

Assignment group

security_engineering

Change type

Normal

Approval required

Yes

Planned start

2026-04-22 18:51:22

Planned end

2026-04-22 20:51:22

Scan ID

titan-banking-demo-20260422T201922Z

Generated at

2026-04-22T18:37:22+00:00

Opened

2026-04-22 18:37:22

Closed

2026-04-22 18:37:22

Close code

—

Attached Security Ticket

AML scoring Databricks workspace has 14 notebooks with hard-coded Snowflake passwords + 3 Unity Catalog tables exposing raw wire data with no row-level security. BSA/AML recordkeeping + FFIEC data-classification policy breach — insider-threat + regulator-visibility risk.

Recommended Fix: databricks secrets put --scope aml-prod --key snowflake-pw && databricks unity-catalog grants update --full-name banking.aml_wires --row-filter amount_filter

AI Close Notes

TITAN CONDUIT opened this high databricks ticket and assigned it to the banking_compliance group for review. STATE: ASSIGNED — awaiting human action. Per TITAN AI policy, configuration changes are NEVER auto-applied and tickets are NEVER auto-closed by TITAN. The assigned group reviews the recommended fix, schedules an approved maintenance window, applies the fix manually, validates via SCOUT rescan, and closes this ticket themselves. TITAN documents and routes — the human owns the change from here.