Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Singapore MAS Technology Risk Management Guidelines expert. Reference-depth framework plugin with scope determination, evidence checklist, and SCF-backed assessment guidance for Singapore-regulated financial institutions.
.claude/skills/grcengclub-sg-mas-trm-expert/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 94% | 0% |
| case-01 | ✗→✓ | ▲ Improved | 0% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 23% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 27% | 0% |
Reference-depth expertise for the Monetary Authority of Singapore (MAS) Technology Risk Management Guidelines, represented in SCF as apac-sgp-mas-trm-2021. This plugin bundles the SCF crosswalk (214 SCF controls to 280 framework controls) with MAS TRM-specific assessment context.
apac-sgp-mas-trm-2021revised January 18, 2021
MAS TRM is Singapore's supervisory guidance for managing technology and cyber risk in financial institutions. It focuses on board and senior management oversight, technology risk governance, system resilience, cyber security, software development, third-party and cloud risk, incident response, and data protection controls. Treat it as a financial-services technology risk framework: the assessor needs to see accountable governance, repeatable operational controls, and evidence that critical systems are resilient and monitored.
MAS TRM applies to financial institutions regulated by MAS, including banks, insurers, capital markets intermediaries, payment service providers, and other regulated entities that depend on technology to deliver financial services. Scope analysis should identify Singapore-regulated entities, critical systems, outsourced service providers, cloud-hosted workloads, customer-facing digital channels, and regional shared-service platforms supporting Singapore operations. Do not treat MAS TRM as a generic ISO 27001 checklist; the Singapore regulatory perimeter and criticality of financial services matter.
Evidence usually centers on board-approved technology risk policies, technology risk registers, critical system inventories, risk assessments, cyber resilience plans, secure SDLC records, vulnerability and penetration test results, privileged access reviews, incident response procedures, outsourcing due diligence, cloud risk assessments, business continuity and disaster recovery test records, and independent assurance reports. MAS TRM is guidance rather than a standalone certification package, so map artifacts to supervisory expectations instead of inventing a formal certificate deliverable.
There is no annual MAS TRM certificate cycle. Assess recurring operating evidence on a risk-based cadence, with stronger scrutiny for critical systems: access reviews, vulnerability management, penetration testing, resilience tests, third-party reviews, and incident exercises should run often enough to support board and senior management oversight. MAS-regulated entities should also align incident escalation and notification evidence with applicable MAS notices, outsourcing requirements, and internal materiality thresholds.
The Monetary Authority of Singapore supervises regulated financial institutions and can use inspections, supervisory findings, directions, penalties, and other regulatory measures when technology risk management is weak. In assessment output, separate MAS TRM control evidence from legal advice about notices, licensing, or penalty exposure.
MAS TRM commonly overlaps with ISO 27001/27002, SOC 2 security and availability criteria, NIST CSF, PCI DSS for payment environments, Singapore PDPA for personal data, and outsourcing/cloud risk requirements. Use the SCF crosswalk for control mechanics, but keep MAS TRM reporting focused on financial institution governance, critical system resilience, cyber operations, secure development, outsourcing, and technology incident readiness.
shared services, and outsourced providers can be in scope when they support Singapore-regulated financial services.
Singapore financial-sector governance, critical-system resilience, outsourcing, and supervisory evidence.
institutions still need due diligence, risk assessment, monitoring, exit planning, and accountability for outsourced or cloud services.
accountability, risk acceptance, and independent assurance are central.
/sg-mas-trm:scope - determine applicability/sg-mas-trm:assess - run a gap assessment/sg-mas-trm:evidence-checklist - enumerate evidence requirementsAll three delegate to /grc-engineer:gap-assessment with SCF framework ID apac-sgp-mas-trm-2021 for the control-by-control mechanics, and wrap the results in MAS TRM-specific terminology.
Full-depth plugins add framework-specific workflow commands tied to the audit ritual. Candidates for this framework:
/sg-mas-trm:critical-system-register - build or review the inventory ofcritical systems, owners, dependencies, RTO/RPO, and resilience evidence.
/sg-mas-trm:outsourcing-cloud-review - assess cloud and outsourced serviceproviders against due diligence, monitoring, exit, and concentration-risk expectations.
/sg-mas-trm:cyber-resilience-check - review monitoring, vulnerability,penetration testing, incident response, and recovery exercise evidence.
/sg-mas-trm:board-pack - summarize technology risk posture and open issuesfor board or senior management oversight.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-03 | pass→fail | 23,655 | 25,579 | +8% | 1 | 1 | 0% | 3,891 | 5,295 | +36% | 0 | 0 | — |
case-04 | fail→pass | 5,536 | 1,832 | -67% | 1 | 1 | 0% | 847 | 1,640 | +94% | 0 | 0 | — |
case-01 | fail→pass | 24,020 | 16,297 | -32% | 1 | 1 | 0% | 3,633 | 3,638 | +0% | 0 | 0 | — |
case-02 | fail→fail | 19,588 | 18,998 | -3% | 1 | 1 | 0% | 3,227 | 4,424 | +37% | 0 | 0 | — |
case-05 | pass→pass | 13,238 | 12,935 | -2% | 1 | 1 | 0% | 2,029 | 3,270 | +61% | 0 | 0 | — |
case-06 | pass→pass | 14,514 | 15,045 | +4% | 1 | 1 | 0% | 2,252 | 3,680 | +63% | 0 | 0 | — |
case-07 | pass→pass | 13,750 | 15,275 | +11% | 1 | 1 | 0% | 2,085 | 3,572 | +71% | 0 | 0 | — |
case-08 | fail→pass | 16,698 | 12,289 | -26% | 1 | 1 | 0% | 2,644 | 3,261 | +23% | 0 | 0 | — |
case-09 | pass→pass | 15,494 | 16,917 | +9% | 1 | 1 | 0% | 2,341 | 3,961 | +69% | 0 | 0 | — |
case-10 | fail→pass | 10,330 | 2,518 | -76% | 1 | 1 | 0% | 1,643 | 1,844 | +12% | 0 | 0 | — |
case-11 | fail→pass | 9,243 | 4,819 | -48% | 1 | 1 | 0% | 1,475 | 1,879 | +27% | 0 | 0 | — |
case-12 | fail→pass | 9,337 | 2,728 | -71% | 1 | 1 | 0% | 1,279 | 1,837 | +44% | 0 | 0 | — |
case-13 | pass→pass | 7,345 | 4,962 | -32% | 1 | 1 | 0% | 1,119 | 2,134 | +91% | 0 | 0 | — |
case-14 | pass→pass | 9,921 | 8,995 | -9% | 1 | 1 | 0% | 1,477 | 2,720 | +84% | 0 | 0 | — |
case-15 | pass→pass | 13,904 | 13,370 | -4% | 1 | 1 | 0% | 2,225 | 3,414 | +53% | 0 | 0 | — |
case-16 | pass→pass | 8,941 | 12,939 | +45% | 1 | 1 | 0% | 1,305 | 3,230 | +148% | 0 | 0 | — |
case-17 | pass→pass | 16,944 | 16,980 | +0% | 1 | 1 | 0% | 2,724 | 4,004 | +47% | 0 | 0 | — |
case-18 | pass→fail | 15,165 | 14,464 | -5% | 1 | 1 | 0% | 2,099 | 3,487 | +66% | 0 | 0 | — |
case-19 | fail→pass | 9,781 | 2,274 | -77% | 1 | 1 | 0% | 1,463 | 1,665 | +14% | 0 | 0 | — |
case-20 | pass→pass | 2,142 | 1,434 | -33% | 1 | 1 | 0% | 270 | 1,559 | +477% | 0 | 0 | — |
case-21 | pass→pass | 19,249 | 20,244 | +5% | 1 | 1 | 0% | 2,821 | 4,439 | +57% | 0 | 0 | — |
case-22 | pass→pass | 14,841 | 18,683 | +26% | 1 | 1 | 0% | 2,303 | 4,284 | +86% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +23 percentage points is the difference between those two pass rates over the 22 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.