用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/hpsgd/turtlestack --skill compliance-audit命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
基于 SOC 职业分类
正在显示 SKILL.md
| name | compliance-audit |
| description | Audit compliance against a regulatory framework — gap analysis, evidence collection, and remediation planning. |
| argument-hint | [framework to audit against, e.g. GDPR, SOC 2, ISO 27001, Privacy Act, Essential Eight] |
| user-invocable | true |
| allowed-tools | Read, Bash, Glob, Grep |
Audit compliance against $ARGUMENTS.
Identify the specific framework, version, and applicable controls:
| Framework | Key reference | Scope |
|---|---|---|
| GDPR | Articles 5–49 | EU personal data processing |
| Australian Privacy Principles | 13 APPs | Australian personal information |
| NZ Privacy Act 2020 | 13 IPPs | NZ personal information |
| SOC 2 | Trust Service Criteria | Service organisation security, availability, processing integrity, confidentiality, privacy |
| ISO 27001 | Annex A controls | Information security management system |
| Essential Eight | 8 strategies, maturity 0–3 | Australian cyber security baseline |
| NZISM | Security manual | NZ government information security |
| Cyber Essentials | 5 controls | UK cyber security baseline |
| HIPAA | Security Rule, Privacy Rule | US health data |
| PCI DSS | 12 requirements | Payment card data |
| NIST CSF 2.0 | 6 functions (Govern, Identify, Protect, Detect, Respond, Recover) | US cybersecurity baseline |
| CIS Controls v8 | 18 prioritised controls | Practical security baseline |
Determine: Which version of the framework? Which controls are applicable given the project's data, users, and jurisdiction? Not every control applies to every system.
Map each applicable framework requirement to an existing control (or lack thereof):
| Requirement ID | Requirement | Current control | Evidence | Status |
|---|---|---|---|---|
| [e.g. GDPR Art. 6] | [description] | [what exists] | [where evidence is] | MET / PARTIAL / GAP |
Evidence types to check:
# Access controls
grep -rn "auth\|authorize\|permission\|role" --include="*.ts" --include="*.py" --include="*.cs"
# Encryption configuration
grep -rn "encrypt\|tls\|ssl\|certificate\|kms" --include="*.yaml" --include="*.yml" --include="*.json" --include="*.tf"
# Logging and audit trails
grep -rn "audit\|log\.\|logger\.\|track\|event" --include="*.ts" --include="*.py" --include="*.cs"
# Data retention/deletion
grep -rn "retention\|purge\|delete\|expire\|ttl" --include="*.ts" --include="*.py" --include="*.cs" --include="*.yaml"
# CI/CD security gates
find . -name "*.yml" -path "*ci*" -o -name "*.yml" -path "*pipeline*" -o -name "*.yml" -path "*workflow*"
For each control marked as MET or PARTIAL, collect verifiable evidence:
| Evidence type | What to look for | Reliability |
|---|---|---|
| Code | Implementation of the control in source code | High — verifiable, versioned |
| Configuration | Infrastructure, CI/CD, or application config | High — verifiable |
| Documentation | Policies, procedures, runbooks | Medium — may be outdated |
| Logs | Audit trails, access logs, change logs | High — if centralised and tamper-resistant |
| Interviews | Team member confirmation of process | Low — self-reported, no verification |
Rules:
For each requirement with status PARTIAL or GAP:
Not all gaps are equal. Rank by regulatory exposure:
| Severity | Criteria | Examples |
|---|---|---|
| Critical | Active non-compliance with mandatory requirement, data currently at risk | No encryption for PII at rest, no breach notification process, no consent mechanism |
| High | Missing control for a high-impact requirement, audit finding likely | Incomplete access controls, no audit trail for data access, no deletion capability |
| Medium | Partial compliance, control exists but has weaknesses | Encryption configured but not verified, logging exists but incomplete, policy exists but unenforced |
| Low | Documentation gap, minor process improvement needed | Policy needs updating, evidence needs better organisation, training records incomplete |
For each gap, define a remediation action:
| Gap ID | Requirement | Severity | Remediation | Owner | Target date | Verification |
|---|---|---|---|---|---|---|
| G-001 | [requirement] | [level] | [specific action] | [person] | [date] | [how to verify completion] |
Prioritisation rules:
# Compliance Audit: [framework] — [scope]
## Audit Summary
- **Framework:** [name and version]
- **Scope:** [systems, data, processes]
- **Date:** [audit date]
- **Auditor:** [who performed this audit]
- **Overall status:** [Compliant / Partially Compliant / Non-Compliant]
## Results Summary
| Status | Count | Percentage |
|---|---|---|
| MET | [n] | [%] |
| PARTIAL | [n] | [%] |
| GAP | [n] | [%] |
| N/A | [n] | [%] |
## Control Matrix
| Req ID | Requirement | Control | Evidence | Status |
|---|---|---|---|---|
| [id] | [description] | [control] | [evidence ref] | MET/PARTIAL/GAP |
## Gap Register
| Gap ID | Requirement | Severity | Description | Exposure |
|---|---|---|---|---|
| G-001 | [req] | Critical/High/Medium/Low | [what is missing] | [what could happen] |
## Remediation Plan
| Gap ID | Remediation | Owner | Target date | Verification |
|---|---|---|---|---|
| G-001 | [action] | [person] | [date] | [how to verify] |
## Re-Audit Schedule
- **Next audit:** [date]
- **Audit triggers:** [regulatory changes, system changes, incidents, 12-month maximum]
- **Evidence refresh:** [how evidence will be kept current between audits]
/grc-lead:risk-assessment — risk-rank gaps found during the compliance audit to prioritise remediation./grc-lead:ai-governance-review — when AI/ML systems are in scope and require specialised governance checks.