Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Security architecture and implementation patterns. Use when designing security controls, implementing authentication/authorization, conducting threat modeling, or ensuring compliance with security frameworks.
.claude/skills/aiskillstore-security-engineering/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✓→✓ | = Same ✓ | 38% | 0% |
| case-03 | ✓→✓ | = Same ✓ | 66% | 0% |
| case-04 | ✓→✓ | = Same ✓ | 39% | 0% |
| case-05 | ✓→✓ | = Same ✓ | 29% | 0% |
| case-06 | ✓→✓ | = Same ✓ | 29% | 0% |
Comprehensive security engineering skill covering application security, infrastructure security, compliance, and incident response.
Layer security controls at multiple levels:
| Layer | Controls | |-------|----------| | Perimeter | Firewall, WAF, DDoS protection | | Network | Segmentation, IDS/IPS, VPN | | Host | Hardening, EDR, patch management | | Application | Input validation, secure coding, SAST/DAST | | Data | Encryption, access control, DLP | | Identity | MFA, SSO, privileged access management |
Core Principles:
Implementation:
Grant Types:
| Grant | Use Case | |-------|----------| | Authorization Code + PKCE | Web/mobile apps | | Client Credentials | Service-to-service | | Device Code | CLI tools, IoT |
Token Best Practices:
Users → Roles → PermissionsBest for: Well-defined organizational hierarchies
If user.department == "engineering" AND
resource.classification == "internal" AND
time.hour BETWEEN 9 AND 17
THEN allowBest for: Complex, dynamic access requirements
Use OPA/Rego or Cedar for externalized policy:
| Risk | Mitigation | |------|------------| | Injection | Parameterized queries, input validation | | Broken Auth | Strong password policy, MFA, rate limiting | | Sensitive Data | Encryption, minimal data collection | | XXE | Disable external entities | | Broken Access | Authorization checks, default deny | | Misconfig | Secure defaults, hardening guides | | XSS | Output encoding, CSP | | Deserialization | Integrity checks, avoid untrusted data | | Components | Dependency scanning, updates | | Logging | Centralized logging, alerting |
SAST (Static Analysis):
DAST (Dynamic Analysis):
Dependency Scanning:
Never:
Do:
| Framework | Focus Area | |-----------|------------| | SOC 2 | Trust services (security, availability, etc.) | | HIPAA | Healthcare data protection | | PCI-DSS | Payment card data | | GDPR | EU personal data protection | | ISO 27001 | Information security management |
| Level | Description | Response Time | |-------|-------------|---------------| | P1 | Active breach, data exfiltration | Immediate | | P2 | Vulnerability being exploited | < 4 hours | | P3 | High-risk vulnerability discovered | < 24 hours | | P4 | Security improvement needed | Next sprint |
references/threat_modeling.md - STRIDE methodology and examplesreferences/compliance_controls.md - Framework-specific control mappings| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 21,000 | 20,344 | -3% | 1 | 1 | 0% | 2,542 | 4,656 | +83% | 0 | 0 | — |
case-02 | pass→pass | 11,969 | 9,208 | -23% | 1 | 1 | 0% | 2,032 | 2,805 | +38% | 0 | 0 | — |
case-03 | pass→pass | 11,304 | 11,104 | -2% | 1 | 1 | 0% | 1,839 | 3,051 | +66% | 0 | 0 | — |
case-04 | pass→pass | 16,958 | 15,761 | -7% | 1 | 1 | 0% | 2,724 | 3,790 | +39% | 0 | 0 | — |
case-05 | pass→pass | 12,488 | 8,162 | -35% | 1 | 1 | 0% | 1,929 | 2,498 | +29% | 0 | 0 | — |
case-06 | pass→pass | 13,657 | 11,544 | -15% | 1 | 1 | 0% | 2,375 | 3,066 | +29% | 0 | 0 | — |
case-07 | pass→pass | 14,925 | 15,475 | +4% | 1 | 1 | 0% | 2,202 | 3,494 | +59% | 0 | 0 | — |
case-08 | pass→pass | 10,896 | 6,667 | -39% | 1 | 1 | 0% | 1,821 | 2,353 | +29% | 0 | 0 | — |
case-09 | pass→pass | 17,892 | 12,315 | -31% | 1 | 1 | 0% | 2,800 | 3,096 | +11% | 0 | 0 | — |
case-10 | pass→pass | 12,984 | 8,317 | -36% | 1 | 1 | 0% | 2,217 | 2,571 | +16% | 0 | 0 | — |
case-11 | pass→pass | 14,111 | 14,482 | +3% | 1 | 1 | 0% | 2,358 | 3,761 | +59% | 0 | 0 | — |
case-12 | pass→pass | 15,206 | 12,658 | -17% | 1 | 1 | 0% | 2,400 | 3,262 | +36% | 0 | 0 | — |
case-13 | pass→pass | 16,039 | 10,208 | -36% | 1 | 1 | 0% | 2,549 | 2,823 | +11% | 0 | 0 | — |
case-14 | pass→pass | 14,745 | 9,939 | -33% | 1 | 1 | 0% | 2,207 | 2,844 | +29% | 0 | 0 | — |
case-15 | pass→pass | 19,399 | 16,217 | -16% | 1 | 1 | 0% | 2,868 | 3,895 | +36% | 0 | 0 | — |
case-16 | pass→pass | 19,040 | 13,560 | -29% | 1 | 1 | 0% | 2,799 | 3,184 | +14% | 0 | 0 | — |
case-17 | pass→pass | 14,112 | 9,003 | -36% | 1 | 1 | 0% | 2,402 | 2,678 | +11% | 0 | 0 | — |
case-18 | pass→pass | 14,897 | 6,772 | -55% | 1 | 1 | 0% | 2,342 | 2,140 | -9% | 0 | 0 | — |
case-19 | pass→pass | 12,293 | 8,721 | -29% | 1 | 1 | 0% | 2,137 | 2,734 | +28% | 0 | 0 | — |
case-20 | pass→pass | 11,927 | 10,483 | -12% | 1 | 1 | 0% | 1,947 | 2,859 | +47% | 0 | 0 | — |
case-21 | fail→fail | 20,874 | 19,467 | -7% | 1 | 1 | 0% | 3,325 | 4,254 | +28% | 0 | 0 | — |
case-22 | fail→fail | 28,391 | 28,367 | -0% | 1 | 1 | 0% | 6,182 | 7,001 | +13% | 0 | 0 | — |
case-23 | fail→fail | 28,037 | 15,610 | -44% | 1 | 1 | 0% | 2,074 | 3,689 | +78% | 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. 23 cases were attempted. The headline lift of 0 percentage points is the difference between those two pass rates over the 23 comparable cases.
Other measured skills in the registry, with their headline benchmark lift.