security-operations

name	security-operations
description	Use this skill when you need SOC (Security Operations Center) setup or management, threat detection and hunting programs, incident response coordination, SIEM configuration and tuning, security monitoring strategy, SRE security integration, or security operations metrics and KPIs. Trigger for active threats, SOC buildout, or operational security program design.
license	Apache-2.0
metadata	{"author":"aviskaar","version":"1.0","tags":["soc","threat-detection","incident-response","siem","threat-hunting","security-operations","monitoring"],"sub-skills":["threat-hunter","incident-responder","sre-operations"]}

Security Operations — VP Security Operations

Role

VP Security Operations owns the 24×7 detection, response, and resilience capability. This skill orchestrates the SOC, threat hunting program, incident response lifecycle, and SRE-security integration to ensure continuous monitoring, rapid detection, and effective containment.

---

Phase 1 — SOC Architecture & SIEM Design

SOC maturity model:

Level	Capability	Description
L1	Alert Triage	Ingest logs, triage alerts, escalate
L2	Investigation	Deep analysis, threat intel correlation
L3	Threat Hunting	Proactive hunt, adversary emulation
L4	Engineering	Detection engineering, toolchain dev
L5	Strategic	Program governance, threat intelligence

SIEM architecture requirements:

• Centralized log ingestion: endpoints, network, cloud, applications, identity
• Log retention: 90 days hot, 365 days warm, 7 years cold (compliance minimum)
• Normalization: CEF/LEEF/ECS format standardization
• Correlation rules: MITRE ATT&CK mapped detections
• Alert deduplication and priority scoring (CVSS × asset criticality)
• Automated enrichment: threat intel feeds (STIX/TAXII), IP reputation, CVE data
• SOAR integration for automated playbook execution

Required log sources (non-negotiable):

Identity:      Active Directory / Entra ID / Okta / IAM
Endpoints:     EDR (CrowdStrike/SentinelOne/Defender)
Network:       Firewall, IDS/IPS, DNS, DHCP, proxy
Cloud:         CloudTrail/Audit Logs (AWS/Azure/GCP)
Applications:  WAF, API gateway, application logs
Email:         O365/Google Workspace security events
Data:          DLP events, database audit logs
Physical:      Badge access, CCTV event integrations

---

Phase 2 — Detection Engineering

Detection rule tiers:

• Tier 1 (Always-on): Known bad IOCs, signature-based, high-fidelity automated block
• Tier 2 (Behavioral): Anomaly-based, ML-assisted, requires L2 validation
• Tier 3 (Hunt-driven): Hypothesis-based, custom KQL/SPL queries from threat-hunter

MITRE ATT&CK coverage targets:

Initial Access:       ≥90% detection coverage
Execution:            ≥85%
Persistence:          ≥80%
Privilege Escalation: ≥90%
Defense Evasion:      ≥70%
Credential Access:    ≥90%
Discovery:            ≥60%
Lateral Movement:     ≥85%
Collection:           ≥75%
Exfiltration:         ≥80%
Command & Control:    ≥85%
Impact:               ≥90%

Alert quality standards:

• False positive rate target: <10% per rule
• Rules exceeding 25% FP → auto-suppressed pending engineering review
• True positive validation: all P1/P2 alerts require post-closure validation

---

Phase 3 — Incident Response Orchestration

Severity classification:

Severity	Definition	Response SLA	Escalation
P1 — Critical	Active breach, data exfiltration, ransomware	15 min acknowledge, 1h contain	CISO + Legal + Exec
P2 — High	Confirmed compromise, insider threat	1h acknowledge, 4h contain	security-operations VP + CISO
P3 — Medium	Suspicious activity, policy violation	4h acknowledge, 24h investigate	L2 SOC
P4 — Low	Informational, compliance flag	24h acknowledge, 72h close	L1 SOC

IR lifecycle (delegate to incident-responder):

1. Detection → Alert fired in SIEM
2. Triage → L1 analyst classifies severity
3. Containment → Automated + manual isolation
4. Eradication → Root cause removal
5. Recovery → Service restoration
6. Post-Incident Review → Lessons learned, detection improvement

Playbook requirements:

• Every P1/P2 scenario must have a documented playbook
• Playbooks reviewed and tabletop-tested annually minimum
• Playbooks cover: ransomware, BEC, data exfiltration, insider threat, DDoS, supply chain compromise, cloud account takeover, AI system compromise

---

Phase 4 — Threat Hunting Program

Delegate to threat-hunter for execution.

Hunting cadence:

• Weekly: IOC sweeps from threat intel feeds
• Bi-weekly: MITRE ATT&CK technique hypothesis hunts
• Monthly: Deep dives on crown jewel systems
• Quarterly: Adversary emulation (purple team)

Hunt hypothesis sources:

• Threat intelligence (ISAC, commercial feeds, OSINT)
• Incident post-mortems from peer organizations
• New MITRE ATT&CK techniques published
• Newly disclosed CVEs affecting environment
• Behavioral anomalies identified by ML

---

Phase 5 — Security Operations Metrics

Operational KPIs (track weekly):

Metric	Target	Critical Threshold
MTTD (Mean Time to Detect)	<1 hour	>4 hours → escalate
MTTR (Mean Time to Respond)	<4 hours	>24 hours → escalate
Alert Volume	Baseline ±20%	>50% spike → investigation
False Positive Rate	<10%	>25% → rule review
P1 Incident Count	0 per month	Any P1 → CISO report
Hunt Coverage (ATT&CK)	≥80% techniques	<60% → gap report
SOC Analyst Utilization	70–85%	>90% → staff review
Playbook Currency	100% reviewed annually	Any expired → immediate

---

Phase 6 — SRE-Security Integration

Delegate to sre-operations for reliability + security fusion.

Integration requirements:

• Security chaos engineering: inject security failures in staging
• Runbook integration: security steps embedded in all SRE runbooks
• SLO/SLA: security event response included in service reliability objectives
• Change management gate: security review required for all production changes (CAB process)
• Blameless post-mortems: security incidents follow SRE post-mortem culture

---

Non-Negotiable SOC Rules

1. 24×7 coverage — no unmonitored windows; follow-the-sun or automated coverage
2. Alert fatigue prevention — never allow >200 daily alerts per analyst
3. Evidence preservation — all IR actions logged and chain-of-custody maintained
4. No silent failures — every alert receives a documented disposition
5. Least-privilege tool access — SOC tools scoped to read-heavy, surgical-write
6. Encrypted communications — all SOC internal comms on encrypted channels
7. Annual IR tabletop mandatory — no exceptions; test all P1 playbooks