Threat Intelligence

Threat Intelligence

Purpose

Manage cyber threat intelligence operations including feed management, IoC/TTP handling, threat hunting, and MITRE ATT&CK mapping. Provide actionable intelligence to SIEM detection rules, SOC operations, and risk management decisions.

Agent Protocol

Trigger

• "threat intelligence", "CTI", "threat feed", "threat intel", "intelligence"
• "IoC", "indicator of compromise", "TTP", "MITRE ATT&CK", "attack chain"
• "threat hunting", "hunt hypothesis", "proactive hunting"
• "intelligence lifecycle", "strategic intelligence", "tactical intelligence"
• "threat actor", "campaign", "TTP mapping", "threat landscape"

Input Context

• Industry vertical (finance, healthcare, tech, government, critical infrastructure)
• Threat model and risk assessment documentation
• Existing intelligence sources (OSINT, commercial feeds, ISACs)
• Current security tools capable of consuming intelligence (SIEM, EDR, firewall, TIP)
• SOC maturity level and threat hunting capability

Output Artifact

Threat intelligence reports (strategic/tactical/operational), IoC lists with context, ATT&CK mappings, hunting hypotheses.

Response Format

## Intelligence Summary
{Relevant threats, actors, campaigns with relevance assessment}

## IoC/TTP Details
{Indicators with context, confidence, MITRE mapping, recommended actions}

## Hunting Recommendations
{Hypotheses, data sources, queries, expected findings}

Completion Criteria

• Intelligence sources identified and prioritized by relevance
• IoC management process defined (collection, validation, enrichment, distribution, retirement)
• TTP mapping to MITRE ATT&CK completed for priority threats
• Threat hunting plan with hypotheses tied to intelligence
• Intelligence sharing and feedback mechanisms established
• CTI reporting cadence and templates defined

Architecture / Decision Trees

Intelligence Source Selection Decision Tree

What is the primary intelligence need?
├── Strategic (long-term trends, threat actor motivations, risk decisions)
│   ├── Industry ISACs (FS-ISAC, Health-ISAC, IT-ISAC)
│   ├── Vendor reports (Mandiant, CrowdStrike, Recorded Future)
│   ├── Government sources (CISA, NCSC, ENISA)
│   └── Open-source intelligence reports (The DFIR Report, VX Underground)
├── Operational (upcoming attacks, campaigns, malware variants)
│   ├── Commercial feeds (Recorded Future, Anomali, ThreatConnect)
│   ├── ISAC threat intelligence feeds
│   ├── MISP sharing communities
│   └── Dark web monitoring (Flashpoint, Digital Shadows)
└── Tactical (IoCs: IPs, domains, hashes, YARA rules)
    ├── Open-source feeds (AlienVault OTX, URLhaus, PhishTank)
    ├── Commercial feeds (VirusTotal, Proofpoint ET Intelligence)
    ├── MISP (self-managed, community shared)
    └── Automated feeds via TIP platform

What is the team's CTI maturity?
├── Level 1 (Initial): No dedicated CTI → Use open-source feeds + vendor reports
├── Level 2 (Defined): Part-time CTI → Add ISAC membership + commercial feeds
├── Level 3 (Managed): Dedicated CTI analyst → Full TIP platform + dark web monitoring
├── Level 4 (Measured): CTI team → Custom intelligence production + threat actor tracking
└── Level 5 (Optimized): Intelligence-driven org → Automated intel-to-detection pipeline

Intelligence Classification Decision Tree

Is the intelligence actionable by existing tools?
├── YES → Can it be automated?
│   ├── YES (IP block, domain sinkhole, hash detection) → Feed to SIEM/EDR/firewall
│   └── NO (requires analyst judgment) → Create CTI note, brief SOC team
└── NO → Is it strategic context?
    ├── YES → Intelligence report for leadership / risk register
    └── NO → Archive for reference, no immediate action

What is the confidence level of the intelligence?
├── HIGH (confirmed by 2+ reliable sources) → Immediate action, automated
├── MEDIUM (single reliable source or multiple moderate sources) → Review before action
└── LOW (unverified, social media, single source) → Monitor only, no action

Is the intelligence relevant to your industry/region/tech stack?
├── YES → Prioritize processing
└── NO → Archive for reference

Workflow

Step 1: Intelligence Requirements Definition

Define Priority Intelligence Requirements (PIRs) aligned to business risk:

priority_intelligence_requirements:
  pir_01:
    question: "Which threat actors are targeting our industry (financial services) in the current quarter?"
    priority: "P1"
    source: "ISAC reports, vendor threat briefs, dark web monitoring"
    consumer: "CISO, risk management, SOC manager"
    update: "Monthly"

  pir_02:
    question: "What are the latest ransomware variants and TTPs affecting our region?"
    priority: "P1"
    source: "Ransomware tracking feeds, The DFIR Report, BleepingComputer"
    consumer: "SOC, detection engineering, incident response"
    update: "Weekly"

  pir_03:
    question: "Are there active campaigns exploiting vulnerabilities in our technology stack?"
    priority: "P2"
    source: "CISA KEV, vendor security advisories, exploit monitoring"
    consumer: "Vulnerability management, detection engineering, IT operations"
    update: "Continuous (feeds) + Weekly summary"

  pir_04:
    question: "What is the current phishing infrastructure targeting our employees?"
    priority: "P2"
    source: "PhishTank, URLhaus, email gateway telemetry"
    consumer: "SOC Tier 1 triage, email security team"
    update: "Continuous (real-time feed)"

  pir_05:
    question: "What are the emerging TTPs for cloud infrastructure attacks?"
    priority: "P3"
    source: "Cloud security research, vendor blogs, incident reports"
    consumer: "Cloud security team, detection engineering"
    update: "Monthly"

Step 2: Intelligence Collection

OSINT Collection Sources:

osint_sources:
  threat_actor_news:
    - "The DFIR Report — detailed incident reports with IOCs"
    - "Mandiant Advantage — threat actor tracking"
    - "CrowdStrike blog — adversary profiles"
    - "VX Underground — malware source code and analysis"
    - "BleepingComputer — ransomware and breach news"
    - "KrebsOnSecurity — investigative security journalism"
    - "SANS ISC — internet storm center daily diaries"

  ioc_feeds:
    - "AlienVault OTX — 100K+ daily IoCs, community curated"
    - "URLhaus — malware distribution URLs (real-time)"
    - "PhishTank — phishing URLs (verified)"
    - "Spamhaus DROP — known malicious IPs"
    - "Abuse.ch Feodo Tracker — C2 server tracking"
    - "CISA AIS — automated indicator sharing (US gov)"
    - "MISP communities — sector-specific sharing"

  exploit_monitoring:
    - "CISA Known Exploited Vulnerabilities (KEV) — actively exploited CVEs"
    - "Exploit-DB — proof of concept exploits"
    - "Metasploit — module updates"
    - "GitHub monitoring — PoC exploit discovery"
    - "Project Zero bug tracker — 90-day disclosure timeline"

  dark_web:
    - "Flashpoint — dark web forum monitoring (commercial)"
    - "Recorded Future — dark web intelligence (commercial)"
    - "Digital Shadows — digital risk monitoring (commercial)"

Automated Collection Pipeline:

import requests
import json
from datetime import datetime, timedelta
from typing import List, Dict

class ThreatIntelCollector:
    def __init__(self, config: Dict):
        self.sources = config['sources']
        self.tip_api = config['tip_api_endpoint']
        self.tip_token = config['tip_api_token']

    def collect_otx(self, api_key: str, days_back: int = 1) -> List[Dict]:
        """Collect pulses from AlienVault OTX."""
        pulses = []
        since = (datetime.utcnow() - timedelta(days=days_back)).isoformat()
        url = f"https://otx.alienvault.com/api/v1/pulses/subscribed"
        headers = {"X-OTX-API-KEY": api_key}

        response = requests.get(url, headers=headers,
                                params={"limit": 50, "modified_since": since})
        if response.status_code == 200:
            data = response.json()
            for pulse in data.get("results", []):
                pulses.append({
                    "source": "alienvault_otx",
                    "id": pulse["id"],
                    "name": pulse["name"],
                    "description": pulse["description"],
                    "tags": pulse.get("tags", []),
                    "tlp": pulse.get("TLP", "amber"),
                    "indicators": [
                        {"type": i["type"], "value": i["indicator"]}
                        for i in pulse.get("indicators", [])
                    ],
                    "collected_at": datetime.utcnow().isoformat(),
                    "reference": pulse.get("reference")
                })
        return pulses

    def collect_cisa_kev(self) -> List[Dict]:
        """Collect Known Exploited Vulnerabilities from CISA."""
        url = "https://www.cisa.gov/sites/default/files/feeds/known_exploited_vulnerabilities.json"
        response = requests.get(url)
        vulns = []
        if response.status_code == 200:
            data = response.json()
            for vuln in data.get("vulnerabilities", []):
                vulns.append({
                    "source": "cisa_kev",
                    "cve_id": vuln["cveID"],
                    "name": vuln["vulnerabilityName"],
                    "date_added": vuln["dateAdded"],
                    "due_date": vuln["dueDate"],
                    "required_action": vuln["requiredAction"],
                    "known_to_be_used_in_ransomware": vuln.get("knownRansomwareCampaignUse", "Unknown"),
                    "collected_at": datetime.utcnow().isoformat()
                })
        return vulns

    def push_to_tip(self, indicators: List[Dict], feed_name: str):
        """Push collected indicators to Threat Intelligence Platform."""
        payload = {
            "feed": feed_name,
            "indicators": indicators,
            "collected_at": datetime.utcnow().isoformat()
        }
        headers = {
            "Authorization": f"Bearer {self.tip_token}",
            "Content-Type": "application/json"
        }
        response = requests.post(f"{self.tip_api}/indicators/bulk", json=payload, headers=headers)
        return response.status_code == 200

    def collect_all(self):
        """Run all collectors."""
        results = {"otx": [], "cisa": [], "errors": []}
        for source_name, source_config in self.sources.items():
            try:
                if source_name == "alienvault_otx":
                    results["otx"] = self.collect_otx(source_config["api_key"])
                elif source_name == "cisa_kev":
                    results["cisa"] = self.collect_cisa_kev()
            except Exception as e:
                results["errors"].append({"source": source_name, "error": str(e)})
        return results

Step 3: Indicator (IoC) Management

IoC Lifecycle:

Collection → Normalization → Deduplication → Enrichment → Scoring → Distribution → Monitoring → Retirement

IoC Scoring and Prioritization:

def score_indicator(indicator: Dict) -> int:
    """Score an indicator from 0-100 based on confidence and relevance."""
    score = 0

    # Source reliability
    source_scores = {
        "cisa": 100, "mandiant": 95, "crowdstrike": 90,
        "alienvault_otx_verified": 70, "urlhaus": 65,
        "phistank_verified": 60, "misp_trusted": 75,
        "alienvault_otx_unverified": 30, "social_media": 10
    }
    score += source_scores.get(indicator.get("source", ""), 20)

    # Age recency (higher for newer)
    age_hours = indicator.get("age_hours", 72)
    if age_hours < 1: score += 30
    elif age_hours < 6: score += 25
    elif age_hours < 24: score += 20
    elif age_hours < 72: score += 15
    else: score += 5

    # Indicator type priority
    type_scores = {
        "C2_domain": 40, "malware_hash": 35, "phishing_url": 30,
        "malicious_ip": 25, "email_address": 20, "registry_key": 15
    }
    score += type_scores.get(indicator.get("type", ""), 10)

    # Relevance to industry
    if indicator.get("industry_relevant"): score += 20
    if indicator.get("tech_stack_relevant"): score += 15

    # Corroboration (multiple sources)
    if indicator.get("source_count", 1) >= 3: score += 25
    elif indicator.get("source_count", 1) >= 2: score += 15

    return min(score, 100)

IoC Distribution Rules:

Step 4: MITRE ATT&CK Mapping

TTP Mapping Template:

threat_actor_profile:
  name: "APT-XX (Fancy Bear variant)"
  aliases: ["APT28", "Sofacy", "Fancy Bear", "Strontium"]
  motivation: "Political espionage, geopolitical advantage"
  target_sectors: ["Government", "Defense", "Think Tanks", "Media"]
  target_regions: ["Europe", "North America", "Eastern Europe"]

  known_ttps:
    initial_access:
      - technique: "T1566.001 — Spearphishing Attachment"
        observed: "Lure documents with malicious macros"
        detection: "Email gateway macro detection, EDR macro execution monitoring"
      - technique: "T1190 — Exploit Public-Facing Application"
        observed: "CVE-2023-XXXX exploitation in VPN appliances"
        detection: "NIDS signatures, vuln scanner, EDR post-exploit"

    execution:
      - technique: "T1204.002 — User Execution: Malicious File"
        observed: "Users double-click weaponized documents"
        detection: "EDR process creation events, User Behavior Analytics"
      - technique: "T1059.001 — PowerShell"
        observed: "PowerShell download cradle for second stage"
        detection: "PowerShell logging (ScriptBlock, Module), EDR command-line capture"

    persistence:
      - technique: "T1547.001 — Registry Run Keys / Startup Folder"
        observed: "HKCU\\Software\\Microsoft\\Windows\\CurrentVersion\\Run"
        detection: "EDR registry monitoring, Sysmon Event ID 13"

    defense_evasion:
      - technique: "T1055.001 — Process Injection: DLL Injection"
        observed: "Inject into explorer.exe or svchost.exe"
        detection: "EDR API call monitoring, memory scanning"
      - technique: "T1027 — Obfuscated Files or Information"
        observed: "Base64-encoded PowerShell, XOR-encrypted payloads"
        detection: "Base64 decode patterns, entropy analysis"

    credential_access:
      - technique: "T1003.001 — LSASS Memory"
        observed: "Mimikatz or comsvcs.dll dump"
        detection: "EDR LSASS access monitoring, Windows Defender Credential Guard"

    lateral_movement:
      - technique: "T1021.006 — Windows Remote Management"
        observed: "WinRM for remote command execution"
        detection: "Windows Event 91 (WinRM), network log analysis"

    command_and_control:
      - technique: "T1071.001 — Web Protocols"
        observed: "HTTPS to compromised legitimate websites"
        detection: "TLS JA3 fingerprinting, domain reputation, beaconing analysis"
      - technique: "T1573.001 — Symmetric Cryptography"
        observed: "Custom XOR or AES encrypted C2 traffic"
        detection: "Traffic entropy analysis, packet timing analysis"

    exfiltration:
      - technique: "T1048.003 — Exfiltration Over Unencrypted Protocol"
        observed: "FTP/SMB exfiltration to staging server"
        detection: "Large outbound file transfers, DLP alerts"

Detection Coverage Scoring:

def calculate_coverage_score(mapped_ttps: Dict, active_detections: List[str]) -> Dict:
    """
    Calculate detection coverage percentage for mapped TTPs.
    Returns coverage stats per tactic.
    """
    tactic_coverage = {}
    total_techniques = 0
    covered_techniques = 0

    for tactic, techniques in mapped_ttps.items():
        total = len(techniques)
        covered = sum(1 for t in techniques if t['technique'] in active_detections)
        tactic_coverage[tactic] = {
            "total": total,
            "covered": covered,
            "coverage_pct": round(covered / total * 100, 1) if total > 0 else 0,
            "techniques": {
                "covered": [t['technique'] for t in techniques if t['technique'] in active_detections],
                "gaps": [t['technique'] for t in techniques if t['technique'] not in active_detections]
            }
        }
        total_techniques += total
        covered_techniques += covered

    return {
        "overall_coverage_pct": round(covered_techniques / total_techniques * 100, 1) if total_techniques > 0 else 0,
        "tactics": tactic_coverage,
        "total_techniques": total_techniques,
        "covered_techniques": covered_techniques,
        "next_priorities": sorted(
            [t for tactics in tactic_coverage.values() for t in tactics['techniques']['gaps']],
            key=lambda t: t.get('priority', 99)
        )[:5]
    }

Step 5: Threat Hunting

Hunt Hypothesis Generation:

hunt_hypotheses:
  h_01:
    hypothesis: "Threat actors may be using PowerShell without -EncodedCommand to avoid detection"
    threat_intel_basis: "Recent campaigns observed using splatted parameters and obfuscated string concatenation"
    data_sources: ["Windows Event 4104 (PowerShell ScriptBlock)", "EDR process command line"]
    query: "index=windows EventCode=4104 ScriptBlockText=* -not ScriptBlockText=*-EncodedCommand* -not ScriptBlockText=*DownloadString* | eval length=len(ScriptBlockText) | where length > 500"
    expected_findings: "Obfuscated PowerShell scripts with unusual parameter passing"
    ttp_mapping: "T1059.001"

  h_02:
    hypothesis: "Adversaries may be using living-off-the-land binaries (LOLBins) for lateral movement"
    threat_intel_basis: "Increasing use of MSBuild, InstallUtil, and Mshta for bypassing application allowlisting"
    data_sources: ["Windows Event 4688 (Process Creation)", "EDR parent-child process tree"]
    query: "index=windows EventCode=4688 (Image=*\\msbuild.exe OR Image=*\\installutil.exe OR Image=*\\mshta.exe) | stats count by ParentImage, CommandLine"
    expected_findings: "LOLBin execution from unusual parent processes (Office, browser, email client)"
    ttp_mapping: "T1218.005, T1218.004, T1218.005"

  h_03:
    hypothesis: "Attackers may be using DNS tunneling for C2 communication"
    threat_intel_basis: "DNS tunneling remains effective and increasingly used by APT groups"
    data_sources: ["DNS logs (query volume, entropy)", "Zeek DNS logs"]
    query: "index=dns | eval domain_length=len(query) | eval subdomain_count=mvcount(split(query, '.')) | eval entropy=... | where domain_length > 30 AND subdomain_count > 5 | stats count by query, src_ip"
    expected_findings: "Unusual DNS query patterns: high frequency, long subdomains, high entropy, TXT queries"
    ttp_mapping: "T1572"

  h_04:
    hypothesis: "Recently disclosed CVE-2026-XXXX may be exploited in our environment"
    threat_intel_basis: "CISA KEV added CVE-2026-XXXX with active exploitation reports"
    data_sources: ["Application logs for affected software", "EDR for post-exploitation indicators", "IDS/IPS signatures"]
    queries:
      - "Search for affected software version in inventory (CMDB, vulnerability scanner)"
      - "Search for exploit-related process creation patterns post-announcement"
      - "Check for new outbound connections from systems running affected software"
    expected_findings: "No exploitation found, or exploitation detected with containment needed"
    ttp_mapping: "T1190"

Hunt Execution Process:

1. Hypothesis: Form clear hypothesis based on threat intelligence
2. Data Collection: Identify and query relevant data sources
3. Analysis: Review results, look for patterns, anomalies, and IOCs
4. Triage: Investigate findings, confirm benign or escalate
5. Documentation: Record findings, update detection rules, feed back to intelligence
6. Metrics: Track hypotheses tested, confirmed findings, new detection rules created

Step 6: Intelligence Reporting

Reporting Cadence:

CTI Report Template:

# THREAT INTELLIGENCE BRIEF — MONTHLY SUMMARY
Date: June 2026
Classification: TLP:AMBER
Prepared by: CTI Team

## Executive Summary
{2-3 sentence overview of the month's most significant threats and their relevance to the organization}

## Industry Threat Landscape
- {Threat actor activity relevant to our industry}
- {New TTPs observed in the wild}
- {Attack campaigns targeting similar organizations}

## Relevant Threat Actors
| Actor | Activity This Month | Relevance | Action Required |
|-------|-------------------|------------|-----------------|
| ACTOR-A | New phishing campaign targeting CFOs | HIGH (finance team targeted) | User awareness training |
| ACTOR-B | Ransomware deployment via VPN vuln | MEDIUM (VPN version patched) | Verify patch compliance |

## Critical Vulnerabilities (CISA KEV)
| CVE | Affected Tech | Impact | Status in Our Environment |
|-----|--------------|--------|-------------------------|
| CVE-2026-XXXX | Product Y | RCE | [Patched / Unpatched / Not applicable] |

## New Detection Rules Created
| Rule Name | ATT&CK Technique | Data Source | Status |
|-----------|-----------------|-------------|--------|
| PowerShell Obfuscation | T1059.001 | Windows Event 4104 | Active |

## IoC Statistics
| Indicator Type | Collected | Reviewed | Distributed | Blocked |
|---------------|-----------|----------|-------------|---------|
| IP Addresses | 1,234 | 800 | 450 | 12 |
| Domains | 567 | 400 | 200 | 5 |
| File Hashes | 890 | 600 | 300 | 8 |
| URLs | 2,100 | 1,500 | 750 | 25 |

## Intelligence Gaps
- {What we don't know about current threats}
- {Recommendations for additional intelligence sources}

## Recommendations
1. {Priority action for SOC/detection engineering}
2. {Priority action for vulnerability management}
3. {Priority action for risk management}

Step 7: Intelligence Sharing and TLP Handling

TLP Classification:

Sharing Platforms:

Step 8: Intelligence-Driven Detection Engineering

Intelligence-to-Detection Pipeline:

Raw Intelligence → Analyze & Enrich → Create Detection → Test → Deploy → Measure → Feedback

Example: CISA KEV to Detection Rule:

Input CISA KEV entry:

{
  "cveID": "CVE-2026-XXXX",
  "vendorProject": "Vendor Y",
  "product": "Product Z",
  "vulnerabilityName": "Product Z Remote Code Execution",
  "dateAdded": "2026-05-28",
  "shortDescription": "Product Z contains an RCE vulnerability in the web interface",
  "requiredAction": "Apply updates per vendor instructions.",
  "knownRansomwareCampaignUse": "Known"
}

Generated detection rules:

# Splunk - Detect exploitation attempts
index=proxy sourcetype=stream:http
| search uri=*ProductZ* AND (uri=*/admin/* OR uri=*/api/*)
| eval threat_actor_cve = "CVE-2026-XXXX"
| eval severity = "critical"
| eval urgency = "high"
| eval recommendation = "Immediately patch Product Z. Investigate for post-exploitation activity."

# Wazuh - Detect vulnerable software version
<rule id="100050" level="10">
  <if_sid>500</if_sid>
  <field name="win.eventdata.product">Product Z</field>
  <field name="win.eventdata.version" type="pcre2">^([0-9]\.|1[0-4]\.)</field>
  <description>CVE-2026-XXXX: Product Z RCE - vulnerable version detected</description>
  <mitre>
    <id>T1190</id>
  </mitre>
</rule>

Common Pitfalls

Pitfall 1: Collecting Everything, Using Nothing

Collecting intelligence from 50+ feeds generates noise and obscures relevant threats. Focus on quality over quantity. Align collection to PIRs. Prune feeds that don't generate actionable intelligence.

Pitfall 2: IoCs Without Context

Distributing IP addresses without context (malware C2 vs phishing URL vs scanner) wastes analyst time. Always include: indicator type, confidence, source, age, associated threat actor, recommended action.

Pitfall 3: Tactical Overemphasis

Focusing only on IoCs misses strategic intelligence. IoCs expire quickly (70% are useful for < 30 days). TTPs provide longer-lasting detection value. Balance tactical (IoCs) with operational (campaigns) and strategic (trends).

Pitfall 4: No Feedback Loop

Intelligence that never gets validated (was the IoC useful? Did we detect anything?) creates blind trust in feeds. Implement feedback: analyst marks IoC as "detected", "false positive", or "not seen". Use feedback to tune feed quality scoring.

Pitfall 5: Ignoring False Positives from Feeds

Some intelligence feeds have 30-50% false positive rates. Without validation, FPs erode trust in CTI. Validate high-severity IoCs before distribution. Use verified feeds for automated actions.

Pitfall 6: Intelligence Not Integrated into Tools

CTI team produces reports but intelligence never reaches SIEM, EDR, or firewall. Automated IoC distribution via TIP platform. Strategic intelligence should influence detection rule priorities.

Pitfall 7: No Threat Hunting Program

Intelligence collected but never used for proactive hunting. Intelligence without hunting is passive. Use each intelligence report to generate at least one hunt hypothesis.

Pitfall 8: One-Size-Fits-All Reporting

The same intelligence report does not serve analysts, managers, and executives. Tailor reports: tactical for SOC, operational for detection engineering, strategic for leadership.

Pitfall 9: Intelligence Hoarding

Not sharing intelligence with peers, ISACs, or CISA. Intelligence sharing improves collective defense. Implement TLP classification to share appropriately. Participate in ISAC intelligence working groups.

Pitfall 10: No Intelligence-Led Testing

Without purple team validation, you don't know if intelligence has improved detection. Use intelligence to drive purple team scenarios. Validate detection rules against real adversary TTPs.

Best Practices

• Define Priority Intelligence Requirements (PIRs) aligned to business risk — collect what you need
• Implement TLP classification for all intelligence products (RED/AMBER/GREEN/CLEAR)
• Use TIP platform (MISP, ThreatConnect, Anomali) for centralized intelligence management
• Automate IoC ingestion from Tier 1 feeds (CISA KEV, AlienVault OTX, MISP)
• Score indicators by confidence, age, source reliability, and relevance before distribution
• Generate detection rules from CISA KEV entries within 24 hours of publication
• Map all intelligence to MITRE ATT&CK for measurable coverage analysis
• Conduct threat hunting based on intelligence — every CTI report should produce at least one hunt hypothesis
• Measure CTI effectiveness: detection rate improvement, intelligence-to-detection latency, feed quality scores
• Participate in ISAC or industry sharing groups for sector-specific intelligence
• Conduct quarterly intelligence gap analysis — what don't we know?
• Maintain TTP/actor profiles for the top 10 threat actors targeting your industry

Performance Considerations

• IoC half-life: 70% of IP-based IoCs are obsolete after 30 days; domain IoCs last 60-90 days; hash-based IoCs last 12+ months
• Feed volume: a single OSINT feed can produce 10K+ IoCs daily. Automated dedup and scoring essential
• Detection engineering: intelligence-to-detection rule should take < 4 hours for critical threats, < 24 hours for high
• MISP performance: 100K events manageable on 4 vCPU, 16GB RAM instance. Scale with Redis caching
• SIEM correlation: threat intel lookups add 10-100ms per event. Cache intel in memory, update hourly
• False positive rate: commercial feeds typically 5-15% FP; OSINT feeds 20-50% FP. Validate before auto-action

Rules

• No intelligence feed consumed without documented source and confidence scoring
• All intelligence products must have TLP classification (RED/AMBER/GREEN/CLEAR)
• IoCs older than 30 days must be reviewed before automated action
• CISA KEV vulnerabilities must be assessed for relevance within 24 hours of publication
• All detection rules based on intelligence must reference the source intelligence
• Hunt hypotheses must be generated from each significant intelligence report
• CTI feedback loop must be closed within 7 days (did the intelligence produce value?)
• Intelligence sharing must comply with TLP classification and legal agreements
• Quarterly intelligence gap analysis must be performed and reported
• Threat actor profiles must be updated within 30 days of significant TTP changes

References

• references/cti-lifecycle.md — Threat Intelligence Lifecycle
• references/osint-collection.md — OSINT Collection
• references/threat-hunting.md — Threat Hunting
• references/threat-intelligence-advanced.md — Threat Intelligence Advanced Topics
• references/threat-intelligence-fundamentals.md — Threat Intelligence Fundamentals
• references/ti-platforms.md — Threat Intelligence Platforms
• references/ti-sharing.md — Threat Intelligence Sharing

Handoff

IoC feeds integrated with siem-engineering for detection rules. TTP mapping informs soc-operations for analyst workflows.