Install any skill in seconds. Free to start, no credit card required.
Get Started Free →ISMS auditing for ISO 27001 compliance, control assessment, and certification support. Use for ISMS audit programs, internal/external ISO 27001 audits, ISO 27002 Annex A control testing, and Stage 1/Stage 2 certification audits.
.claude/skills/borghei-isms-audit-expert/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 213% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 91% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 186% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 291% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 356% | 0% |
Internal and external ISMS audit management for ISO 27001 compliance verification, security control assessment, and certification support.
Before planning or executing the audit, confirm these inputs. If any is unknown or vague, ASK — do not assume:
Stop rule: ask only the 2-3 that most change the output. If the user says "just draft it," proceed and list your assumptions at the top of the audit plan.
| Risk Level | Audit Frequency | Examples | |------------|-----------------|----------| | Critical | Quarterly | Privileged access, vulnerability management, logging | | High | Semi-annual | Access control, incident response, encryption | | Medium | Annual | Policies, awareness training, physical security | | Low | Annual | Documentation, asset inventory |
ISMS AUDIT PLAN 2026
Prepared by: Information Security Manager
Approved by: CISO
Date: 2026-01-15
Q1 2026 (January-March)
Scope: Privileged access (A.8.2, A.8.18), Logging (A.8.15, A.8.16)
Auditor: External consultant (independence required)
Risk level: Critical
Q2 2026 (April-June)
Scope: Access control (A.8.3-A.8.5), Incident response (A.5.24-A.5.28)
Auditor: Internal audit team
Risk level: High
Q3 2026 (July-September)
Scope: Physical security (A.7.1-A.7.14), HR security (A.6.1-A.6.8)
Auditor: Internal audit team
Risk level: Medium
Q4 2026 (October-December)
Scope: Policies (A.5.1-A.5.8), Asset management (A.5.9-A.5.14)
Auditor: Internal audit team
Risk level: Medium-Low
Coverage: 93/93 Annex A controls scheduled across 4 quarters| Method | Use Case | Example | |--------|----------|---------| | Inquiry | Process understanding | Interview Security Manager about incident response | | Observation | Operational verification | Watch visitor sign-in process at reception | | Inspection | Documentation review | Check access approval records for last quarter | | Re-performance | Control testing | Attempt login with weak password to verify policy enforcement |
Organizational Controls (A.5): Information security policies, roles and responsibilities, segregation of duties, contact with authorities, threat intelligence, information security in projects.
People Controls (A.6): Screening and background checks, employment terms, security awareness and training, disciplinary process, remote working security.
Physical Controls (A.7): Physical security perimeters, entry controls, securing offices and facilities, physical security monitoring, equipment protection.
Technological Controls (A.8): User endpoint devices, privileged access rights, access restriction, secure authentication, malware protection, vulnerability management, backup and recovery, logging and monitoring, network security, cryptography.
CONTROL TEST WORKING PAPER
Control: A.8.2 - Privileged access rights
Objective: Privileged access is restricted and managed
Test date: 2026-03-10
Auditor: J. Smith
Test procedure:
1. Obtained list of privileged accounts from IAM system (42 accounts)
2. Selected sample of 10 accounts (25% sample rate)
3. For each account, verified:
- Documented business justification exists
- Manager approval on file
- Quarterly access review completed
- No dormant accounts (last login within 90 days)
Results:
- 8/10 accounts: All criteria met (PASS)
- 1/10: Missing quarterly review for Q4 2025 (MINOR NC)
- 1/10: No documented business justification (MINOR NC)
Conclusion: Control partially effective - minor nonconformity raised
Finding reference: ISMS-2026-007| Severity | Definition | Response Time | |----------|------------|---------------| | Major Nonconformity | Control failure creating significant risk | 30 days | | Minor Nonconformity | Isolated deviation with limited impact | 90 days | | Observation | Improvement opportunity | Next audit cycle |
Finding ID: ISMS-2026-007
Control Reference: A.8.2 - Privileged access rights
Severity: Minor Nonconformity
Evidence:
- 1 of 10 sampled privileged accounts missing Q4 2025 review
- 1 of 10 sampled accounts lacks documented business justification
- Screenshots of IAM records and review log exported 2026-03-10
Risk Impact:
- Unreviewed privileged access increases insider threat exposure
- Non-justified accounts may represent unnecessary attack surface
Root Cause:
- Access review process relies on manual tracking; no automated reminder
Recommendation:
- Implement automated quarterly review reminders via IAM platform
- Require business justification field as mandatory in provisioning workflow
- Backfill missing reviews within 14 days| Period | Focus | |--------|-------| | Year 1, Q2 | High-risk controls, Stage 2 findings follow-up | | Year 1, Q4 | Continual improvement, control sample | | Year 2, Q2 | Full surveillance | | Year 2, Q4 | Re-certification preparation |
| Script | Purpose | Usage | |--------|---------|-------| | isms_audit_scheduler.py | Generate risk-based audit plans | python scripts/isms_audit_scheduler.py --year 2026 --format markdown |
bash# Generate annual audit plan python scripts/isms_audit_scheduler.py --year 2026 --output audit_plan.json # With custom control risk ratings python scripts/isms_audit_scheduler.py --controls controls.csv --format markdown # Generate plan for specific quarters only python scripts/isms_audit_scheduler.py --year 2026 --quarters Q1 Q2 --format json
| File | Content | |------|---------| | iso27001-audit-methodology.md | Audit program structure, pre-audit phase, certification support | | security-control-testing.md | Technical verification procedures for ISO 27002 controls | | cloud-security-audit.md | Cloud provider assessment, configuration security, IAM review |
| KPI | Target | Measurement | |-----|--------|-------------| | Audit plan completion | 100% | Audits completed vs. planned | | Finding closure rate | >90% within SLA | Closed on time vs. total | | Major nonconformities | 0 at certification | Count per certification cycle | | Audit effectiveness | Incidents prevented | Security improvements implemented |
| Framework | ISMS Audit Relevance | |-----------|---------------------| | GDPR | A.5.34 Privacy, A.8.10 Information deletion | | HIPAA | Access controls, audit logging, encryption | | PCI DSS | Network security, access control, monitoring | | SOC 2 | Trust Services Criteria mapped to ISO 27002 |
| Problem | Possible Cause | Resolution | |---------|---------------|------------| | Audit plan does not cover all 93 Annex A controls within the certification cycle | Controls not inventoried against the 2022 four-theme structure or risk-based scheduling gaps | Use isms_audit_scheduler.py with a complete controls CSV covering all 93 controls; ensure the 3-year cycle allocates quarterly audits for critical controls and annual coverage for all others | | Major nonconformity found during certification audit | Systemic control failure or complete absence of a required ISMS element | Conduct immediate root cause analysis; develop corrective action plan with 30-day target; re-test the control with fresh evidence; schedule verification audit with certification body | | Auditor independence challenged by certification body | Internal auditors assigned to areas they manage or operate | Establish clear auditor independence policy; never assign auditors to areas they are responsible for; consider external consultants for high-risk control areas; document independence verification for each audit | | Evidence collection incomplete for technological controls (A.8) | Technical configurations not captured, logs not retained, or screenshots not timestamped | Prepare control-specific evidence checklists before audit; request system administrators to export configurations; ensure log retention covers the audit period; timestamp all evidence artifacts | | Finding closure rate below 90% target | Corrective actions not prioritized, unclear ownership, or insufficient follow-up | Assign specific owners with due dates for every finding; implement automated tracking with escalation at 50% and 75% of SLA; conduct monthly corrective action reviews | | Surveillance audit identifies regression in previously passed controls | Controls degraded after initial certification due to staff changes, system updates, or process drift | Implement continuous compliance monitoring (not just annual checks); schedule monthly control spot-checks for high-risk areas; include control effectiveness in management review | | Sample-based testing misses systemic issues | Sample size too small or selection biased toward known-good records | Calculate sample size based on population and risk level (minimum 25 for quarterly reviews); use random selection methods; increase sample for areas with prior findings |
In Scope:
Out of Scope:
infrastructure-compliance-auditor for technical checksinformation-security-manager-iso27001 for implementation guidanceImportant Notes:
| Skill | Integration | When to Use | |-------|-------------|-------------| | information-security-manager-iso27001 | ISMS implementation provides the controls and documentation that audits assess | When audit findings require control improvements or ISMS enhancements | | infrastructure-compliance-auditor | Technical infrastructure checks provide audit evidence for Annex A technological controls | When audit requires evidence of A.8 technological control implementation | | soc2-compliance-expert | SOC 2 audit evidence and Trust Services Criteria overlap with ISO 27001 controls | When organization maintains both ISO 27001 and SOC 2 compliance programs | | capa-officer | Audit findings requiring formal corrective action feed into CAPA process | When major nonconformities require structured root cause analysis and corrective action |
Generates risk-based annual audit plans with quarterly scheduling based on control risk ratings.
| Flag | Required | Description | |------|----------|-------------| | --year <year> | No | Target year for audit plan (default: current year) | | --controls <file> | No | CSV file with custom control risk ratings (columns: control_id, name, risk); defaults to built-in risk ratings for 18 key controls | | --quarters <list> | No | Generate plan for specific quarters only (e.g., --quarters Q1 Q2) | | --format <fmt> | No | Output format: json (default) or markdown | | --output <file> | No | Export audit plan to specified file path |
Audit Frequency by Risk Level:
critical: Quarterly (4x per year) -- e.g., A.8.2 Privileged access, A.8.5 Authentication, A.8.8 Vulnerabilities, A.8.15 Logginghigh: Semi-annual (2x per year) -- e.g., A.5.15 Access control, A.5.24 Incident management, A.8.7 Malware protectionmedium: Annual (1x per year) -- e.g., A.5.1 Policies, A.6.3 Awareness training, A.7.1 Physical perimeterslow: Annual (1x per year) -- e.g., Documentation, asset inventoryOutput: Quarterly audit schedule with control assignments, auditor allocation guidance, risk-based prioritization, and coverage tracking ensuring all controls are scheduled within the certification cycle.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | fail→pass | 14,107 | 14,087 | -0% | 1 | 1 | 0% | 2,113 | 6,616 | +213% | 0 | 0 | — |
case-01 | fail→fail | 19,653 | 16,999 | -14% | 1 | 1 | 0% | 3,530 | 7,789 | +121% | 0 | 0 | — |
case-02 | fail→fail | 22,155 | 19,655 | -11% | 1 | 1 | 0% | 3,824 | 7,824 | +105% | 0 | 0 | — |
case-03 | fail→pass | 31,165 | 30,767 | -1% | 1 | 1 | 0% | 5,299 | 10,135 | +91% | 0 | 0 | — |
case-05 | fail→pass | 13,194 | 7,803 | -41% | 1 | 1 | 0% | 2,008 | 5,739 | +186% | 0 | 0 | — |
case-06 | fail→pass | 8,979 | 5,599 | -38% | 1 | 1 | 0% | 1,351 | 5,284 | +291% | 0 | 0 | — |
case-07 | pass→pass | 13,863 | 10,605 | -24% | 1 | 1 | 0% | 2,324 | 6,123 | +163% | 0 | 0 | — |
case-08 | pass→pass | 10,953 | 6,354 | -42% | 1 | 1 | 0% | 1,664 | 5,421 | +226% | 0 | 0 | — |
case-09 | fail→pass | 7,058 | 2,785 | -61% | 1 | 1 | 0% | 1,058 | 4,820 | +356% | 0 | 0 | — |
case-10 | fail→pass | 12,390 | 14,754 | +19% | 1 | 1 | 0% | 2,061 | 6,815 | +231% | 0 | 0 | — |
case-11 | pass→pass | 9,807 | 8,191 | -16% | 1 | 1 | 0% | 1,554 | 5,674 | +265% | 0 | 0 | — |
case-12 | pass→pass | 5,291 | 3,077 | -42% | 1 | 1 | 0% | 878 | 4,877 | +455% | 0 | 0 | — |
case-13 | pass→pass | 7,837 | 4,407 | -44% | 1 | 1 | 0% | 1,398 | 5,001 | +258% | 0 | 0 | — |
case-14 | fail→pass | 4,726 | 4,555 | -4% | 1 | 1 | 0% | 865 | 5,194 | +500% | 0 | 0 | — |
case-15 | pass→pass | 3,508 | 4,592 | +31% | 1 | 1 | 0% | 658 | 5,143 | +682% | 0 | 0 | — |
case-16 | pass→pass | 16,223 | 11,629 | -28% | 1 | 1 | 0% | 2,778 | 6,366 | +129% | 0 | 0 | — |
case-17 | pass→pass | 5,034 | 5,301 | +5% | 1 | 1 | 0% | 841 | 5,132 | +510% | 0 | 0 | — |
case-18 | pass→pass | 11,557 | 10,315 | -11% | 1 | 1 | 0% | 1,794 | 5,976 | +233% | 0 | 0 | — |
case-19 | pass→fail | 5,745 | 8,422 | +47% | 1 | 1 | 0% | 620 | 5,048 | +714% | 0 | 0 | — |
case-20 | fail→fail | 28,056 | 34,223 | +22% | 1 | 1 | 0% | 4,918 | 10,564 | +115% | 0 | 0 | — |
case-21 | fail→pass | 13,781 | 13,449 | -2% | 1 | 1 | 0% | 2,337 | 6,571 | +181% | 0 | 0 | — |
case-22 | fail→pass | 25,196 | 28,685 | +14% | 1 | 1 | 0% | 4,091 | 9,203 | +125% | 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 +36 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is 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.