DEV Community

Davi
Davi

Posted on Originally published at blog.mago.team

Threat Intelligence 101: From Raw Data to Actionable Insights

The SOC got the alert at 2 a.m. By the time the analyst correlated the IOC with the logs and confirmed relevance, the attacker had already moved laterally into three internal servers. The intelligence was right. The pipeline was the problem.

Most organizations already pay for threat intelligence feeds. The gap is not data: it is operationalization. IOCs arrive in dashboards and PDFs that never automatically translate into detection rules, blocklists, or playbooks. The speed advantage that intelligence should deliver gets destroyed by manual correlation that no team can sustain at scale.

The Operationalization Gap

Organizations that suffered breaches in 2024 and 2025 were not operating without intelligence. They had feed subscriptions. The pattern documented in security research is consistent: intelligence arrives, operations cannot absorb it within the relevant time window, the moment passes, and the lesson disappears without a record.

An IOC that fires without an associated playbook is a documented detection gap. CISA demonstrated with the AIS (Automated Indicator Sharing) program that machine-speed sharing is a solved problem: STIX 2.1 for structure, TAXII 2.1 for transport. What AIS does not solve is the last mile: mapping indicators to SIEM rules and EDR actions, a step that most teams still handle manually.

Automating the conversion of CTI reports into detection rules exists as an active research field, but adoption in SOC teams has not kept pace with the volume of intelligence produced. Organizations continue paying for intelligence that never reaches the detector.

What Threat Intelligence Actually Is

Intelligence is not the same as data. STIX 2.1, ratified by OASIS in June 2021, defines structured objects with 18 domain objects and 18 cyber-observable objects to represent context, relationships, and relevance. A CSV list of IPs has none of that.

The practical distinction splits intelligence into three layers: strategic informs long-term decisions, operational supports campaign planning, and tactical provides IOCs for immediate detection. Most commercial feeds deliver exclusively the tactical layer, which also has the shortest shelf life.

MITRE ATT&CK v18 (October 2025) catalogs 14 tactics, 216 techniques, and 475 sub-techniques for the Enterprise environment. Adversary techniques change on a quarterly timescale. IOCs change on an hourly timescale. Building detections based on ATT&CK TTPs produces rules that survive attacker infrastructure rotation.

Commercial Feeds Solve the Wrong Problem

Commercial feeds deliver what is easy to deliver: IP addresses, file hashes, and domains. The problem is that attackers rotate IPs within hours, and half of fresh feed indicators generate false positives from legitimate business traffic, requiring manual triage before any application. With each new batch of indicators, analysts must filter before applying, destroying the automation gain.

The result is context loss at delivery: the SOC receives values, not intelligence. An IP address without information about which campaign, which actor, and which technique it represents is noise dressed as signal.

IP blocklists have a shelf life of 24 to 72 hours under normal conditions, and shorter when the attacker notices they are being blocked. Relying exclusively on tactical feeds anchors detection in infrastructure-centric thinking, while the approach that survives is behavior-centric. ATT&CK was built precisely for that paradigm shift.

Underused Passive Sources: What You Already Have

Operationalizing threat intelligence does not require depending exclusively on external feeds: the organization's own attack surface already generates primary intelligence that no vendor can see. Certificate Transparency logs are public and free. When an attacker registers a typosquat domain for a phishing campaign, they need an SSL certificate before launching the infrastructure. That certificate gets logged in CT logs and is available for query before the attack starts. IOC feeds only capture the domain after the campaign is active and observed by third parties.

SecYbers documents that CT logs combined with DNS and WHOIS data enable full adversary infrastructure mapping before the active phase of the campaign. Subdomain drift, the creation of unmanaged subdomains on the organization's own domains, exposes attack surface that no external feed can see because it is internal. Technology fingerprinting reveals exposed services before attackers discover them through their own scans.

These sources are continuous intelligence generated by the organization's own attack surface. Tools like intel.mago.team/spells monitor CT logs and infrastructure detection automatically as intelligence sources; for point-in-time queries, crt.sh offers free search and SSLMate's Cert Spotter provides continuous monitoring with a free tier. Most teams outsource intelligence generation to external feeds while already operating high-quality primary sources without processing them.

How Automation Closes the Gap

The minimum functional pipeline has four stages: ingestion of structured sources, enrichment with context, mapping to ATT&CK TTPs, and automatic delivery to SIEM or EDR. The gap lives between mapping and delivery, which is exactly where manual work intervenes and destroys the speed that automation should provide.

TAXII 2.1 solves the transport layer: feeds that publish via TAXII allow automated ingestion without manual CSV parsing. STIX 2.1 solves structure: typed objects with explicit relationships let downstream systems process indicators without human intervention. CISA provides AIS access via TAXII to any organization that wants to start with government-sourced intelligence.

Research published on arXiv in 2026 (From IOCs to Regex, arXiv:2604.12228) demonstrates that LLMs can convert CTI reports into regex expressions for detection with a significant degree of automation. The problem is not technological: teams that do not define what they need to detect before buying feeds have no way to evaluate what to automate.

Building the Minimum Pipeline

The most common mistake is starting with the feed and working back to requirements. The functional pipeline reverses that order.

Step one is defining intelligence requirements based on realistic threats for the organization's sector, size, and technology stack. Without that, any feed generates noise. Step two is an attack surface inventory, because you cannot prioritize intelligence without knowing what to protect. Step three is structured ingestion: feeds in STIX format via TAXII enable automation; raw CSVs do not. Step four is TTP mapping, building detections based on adversary behavior rather than point-in-time indicators that expire. Step five is the feedback loop: measure how many alerts generated by last quarter's intelligence produced an actual security decision change.

The 2025 arXiv research on MITRE ATT&CK applications (arXiv:2502.10825) identifies automated TTP extraction from reports as a critical gap. Most organizations perform that mapping manually, making the most time-consuming pipeline stage also the most error-prone.

What to Operationalize First

Operationalization priority follows immediate impact with minimal configuration friction.

Credential exposure monitoring detects when organizational accounts appear in public dumps before attackers use them. Certificate issuance monitoring via CT logs detects phishing infrastructure before campaign launch. Subdomain drift detection exposes attack surface that was never inventoried.

All three share one property: they produce actionable alerts, not reports. The output is a suspect domain to block, a credential to rotate, a subdomain to disable. Define your requirements before buying the next feed. Set up CT log monitoring for your domains. Map your detections to ATT&CK TTPs, not IP lists. Intelligence that never reaches the detector is indistinguishable from having no intelligence.

Top comments (0)