Full pre-production readiness audit combining all domain audits. Use when asked to assess production readiness, do a launch checklist, or perform a go/no-go review. Covers OWASP Top 10, GDPR, GPSR, multi-tenancy, and operational readiness.
Installs into .claude/skills of the current project.
Are you the author of Audit Preproduction?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/lroy-stack-audit-preproduction)
---
name: Audit Pre-Production
description: >
Full pre-production readiness audit combining all domain audits. Use when asked to assess
production readiness, do a launch checklist, or perform a go/no-go review. Covers OWASP
Top 10, GDPR, GPSR, multi-tenancy, and operational readiness.
---
# Audit Pre-Production
Master audit that orchestrates all domain audits and produces a go/no-go decision.
## Prerequisites
- Run (or have results from) all 6 domain audits:
- `audit-frontend`, `audit-admin`, `audit-database`, `audit-api`, `audit-podclaw`, `audit-infrastructure`
- Read existing audit reports at workspace root (`AUDIT_*_[DATE].md`)
- Read `CLAUDE.md` for architecture overview
## Workflow
### Phase 1: Aggregate Domain Audits
1. **Collect existing reports**:
- Read all `AUDIT_*_[DATE].md` files
- Extract CRITICAL and FAIL items from each
- Build consolidated issue list
2. **Run missing audits**:
- If any domain audit is missing or stale (>7 days), trigger it
- Use the corresponding audit skill
### Phase 2: OWASP Top 10 Cross-Check
3. **A01 Broken Access Control**:
- RLS policies enforced? (from DB audit)
- Auth on every endpoint? (from API audit)
- Admin RBAC working? (from admin audit)
4. **A02 Cryptographic Failures**:
- Secrets in env vars, not code? (from infra audit)
- HTTPS enforced? (from infra audit)
- Password hashing algorithm? (from admin audit)
5. **A03 Injection**:
- SQL injection vectors? (from API audit)
- XSS vectors? (from frontend audit)
- Command injection? (from API/PodClaw audit)
6. **A04 Insecure Design**:
- Multi-tenancy isolation verified? (from DB audit)
- Chat isolation per user? (from frontend audit)
- Agent sandboxing? (from PodClaw audit)
7. **A05 Security Misconfiguration**:
- Security headers present? (from frontend/admin/infra audit)
- Default credentials removed? (from admin audit)
- Debug mode disabled? (from all audits)
8. **A06 Vulnerable Components**:
- `npm audit` results for frontend and admin
- Python dependency audit for PodClaw
- Docker image CVE scan
9. **A07 Authentication Failures**:
- Brute force protection? (from admin audit)
- Session management? (from frontend/admin audit)
- MFA available? (from admin audit)
10. **A08 Data Integrity Failures**:
- Webhook signature validation? (from API audit)
- CI/CD pipeline security? (from infra audit)
- Dependency integrity? (lockfile pinning)
11. **A09 Logging & Monitoring**:
- Audit logging? (from admin audit)
- Agent activity logging? (from PodClaw audit)
- Infrastructure monitoring? (from infra audit)
12. **A10 SSRF**:
- Can users trigger server-side requests? (AI calls, image URLs)
- Are URLs validated before fetching?
- Is there an allowlist for external domains?
### Phase 3: Regulatory Compliance
13. **GDPR compliance**:
- Is there a privacy policy?
- Can users export their data?
- Can users delete their account and all data?
- Is consent tracked for marketing communications?
- Are data processing records maintained?
14. **GPSR compliance** (EU Product Safety):
- Do ALL products have safety_information in product_details?
- Is manufacturer info present?
- Are material compositions listed?
- Is care instruction info available?
15. **Cookie consent**:
- Is there a cookie banner?
- Are non-essential cookies blocked until consent?
- Is consent revocable?
### Phase 4: Operational Readiness
16. **Deployment checklist**:
- Can the system be deployed from scratch with `start.sh`?
- Is there a rollback procedure?
- Are backups configured and tested?
- Is there a runbook for common failures?
17. **Monitoring & alerting**:
- Are health checks configured for all services?
- Is there uptime monitoring?
- Are error alerts configured? (email, Telegram, etc.)
- Is there a status page?
18. **Performance baseline**:
- Are page load times acceptable? (<3s on 3G)
- Are API response times reasonable? (<500ms p95)
- Is there a CDN for static assets?
- Are images optimized?
### Phase 5: Go/No-Go Decision
19. **Classify all findings**:
- **BLOCKER**: Must fix before launch (any CRITICAL from domain audits)
- **SHOULD FIX**: Fix within first week post-launch
- **NICE TO HAVE**: Backlog items
20. **Decision matrix**:
- 0 BLOCKERS → GO
- 1+ BLOCKERS → NO-GO (list required fixes)
## Output Format
Generate `AUDIT_PREPRODUCTION_[DATE].md` at workspace root with:
```markdown
# Pre-Production Audit — [DATE]
## GO / NO-GO Decision: [GO | NO-GO]
## Blockers (must fix before launch)
[List with source audit and fix description]
## OWASP Top 10 Matrix
| Category | Status | Source | Notes |
|---|---|---|---|
## GDPR Compliance
| Requirement | Status | Notes |
|---|---|---|
## GPSR Compliance
| Requirement | Status | Notes |
|---|---|---|
## Should Fix (first week)
[Prioritized list]
## Nice to Have (backlog)
[Deprioritized list]
## Domain Audit Summary
| Audit | Date | CRITICAL | FAIL | WARN | PASS |
|---|---|---|---|---|---|
```