◆ CODEPREFLIGHT
COBOL Assessment Report
UK Clearing Bank risk analysis — assessment only, no transpilation performed
Risk Categories (CB-CAT-01 – CB-CAT-16)
| Code | Category | Max Severity | Findings |
| CB-CAT-01 |
Identification and Programme Structure |
INFO
|
1 |
| CB-CAT-02 |
Date and Time Handling |
P2
|
1 |
| CB-CAT-03 |
Copybook and Record Layout |
P2
|
1 |
| CB-CAT-04 |
Numeric Precision and Arithmetic |
P3
|
1 |
| CB-CAT-05 |
File I/O and VSAM Operations |
P1
|
1 |
| CB-CAT-06 |
Regulatory Reporting Integrity |
P3
|
1 |
| CB-CAT-07 |
Settlement Finality and Payment Integrity |
P3
|
1 |
| CB-CAT-08 |
Image Processing and Cheque Clearing |
P2
|
1 |
| CB-CAT-09 |
Real-Time Payment Concurrency and Race Conditions |
P4
|
1 |
| CB-CAT-10 |
AML and Sanctions Compliance |
P3
|
1 |
| CB-CAT-11 |
Batch Processing Integrity |
P2
|
1 |
| CB-CAT-12 |
Working Storage and Data Structures |
INFO
|
1 |
| CB-CAT-13 |
Error Handling and Recovery |
P4
|
1 |
| CB-CAT-14 |
Standing Order and Mandate Management |
P2
|
1 |
| CB-CAT-15 |
Consumer Duty and Conduct Risk |
P3
|
1 |
| CB-CAT-16 |
Database Integrity and Audit Trail |
P4
|
1 |
Executive Summary
This assessment identified 16 findings: 1 at P1 (immediate priority), 5 at P2 (near-term attention), 5 at P3 (scheduled review), 3 at P4 (noted for awareness), 1 informational.
GEOSURV is a sequential batch programme written in 1987 for the National Geological Survey, running on IBM-370 hardware with no CICS, DB2, or MQ dependency. It processes field sample records through classification, validation, averaging, and catalogue generation. The programme is not a UK clearing bank payment system and falls entirely outside the scope of the CB-CAT taxonomy's payment rail, regulatory reporting, and financial crime categories; this analysis therefore applies the structural and data-integrity portions of that taxonomy to the programme as presented. Fourteen findings have been identified across the severity range. The most significant risks are: a subscript in paragraph 441-TALLY-MINERALS that will cause a table boundary violation and programme abend on any mineral code outside the range 1–15; an incorrect average latitude calculation that accumulates raw latitude values without dividing by sample count; unrestricted string construction that will silently overwrite CATALOG-REC data when concatenated fields exceed 80 characters; and missing FILE STATUS checks on all three files. No regulatory consequences apply; all risks are operational.
---
Full structured assessment with evidence, detail and action items in cobol_assessment.txt