Skip to main content

CTI as a Code in Practice: Proactive Threat Assessment — CelltronX Telecom

A complete walkthrough of the proactive methodology: how four intelligence triggers, a contractor supply chain threat, and an INCD compliance gap become a sprint-ready detection backlog.

All organizations, names, and data are fictional. This is training assignment A02 from the CTI as a Code repository.


Contents


The Scenario

CelltronX Ltd. is an Israeli telecommunications provider with 4.2 million mobile subscribers. It operates SS7 signaling infrastructure, a network management system (ENM) for routing control, and billing systems handling 8 million subscriber records per month. Three months ago it was awarded a $210 million government contract to provide connectivity for the Israeli Prime Minister's Office — a fact disclosed on TASE (the Tel Aviv Stock Exchange).

There has been no incident. The SOC has received no alerts. Nobody has called the CTI team with a problem.

But four intelligence signals have arrived in the past three weeks. The CISO has commissioned a threat assessment.


Step 0: Intake — Capturing the Commission Before the Analysis

The assessment begins with an intake call, not with opening a threat report. The intake establishes what the analysis must answer, who it must serve, and what constraints exist.

# Proactive Assessment Intake — ASSESS-2024-002 — 2024-10-15

Completed by: CTI Lead
Commissioned by: Dr. Rotem Katz, CISO (hired 90 days ago; Board mandate:
"Demonstrate measurable improvement in security posture within 6 months")

## Assessment trigger
CERT-IL TLP:AMBER bulletin received 10 days ago. Peer company incident 3 weeks ago.

## What questions must this assessment answer?
- Is CelltronX currently targeted by the same actor that hit MobileTech IL?
- What is the realistic attack path from the contractor access to our crown jewels?
- What detection capabilities do we need to deploy in the next 30 days?

## Expected deliverable
Technical brief for security team + executive brief for CISO/Board +
detection backlog with prioritization.

## Hard deadline
INCD-CID pre-audit is in 6 months. Board review in 4 weeks.

## Engineering capacity for detection backlog
Detection team: 3 analysts, 6–8 rules per 30-day sprint.
Currently backlogged: 14 pending rule requests.

## Analyst assessment (initial)
Threat relevance: HIGH. Same contractor type as confirmed victim.
Top 3 risks: (1) ENM access via contractor, (2) billing API credential abuse,
(3) PM Office connectivity disruption.

The engineering capacity constraint (6–8 rules per sprint) is captured in intake because it directly determines the scope of the detection backlog. Writing 30 rules for a team that can deploy 8 in a month wastes analysis time. The backlog must be realistic from day one.

git init && git add .
git commit -m "ASSESS-002: scaffold initialized"

git add 00-scope/intake-2024-10-15.md
git commit -m "ASSESS-002: intake — 4 triggers, INCD audit in 6mo, 6-8 rules/sprint capacity"

Step P3: Trigger Intelligence Assessment — Four Signals, One Picture

A proactive assessment is triggered by intelligence, not by an incident. The first analytical step is to assess each trigger rigorously: what is the source reliability, what is the information reliability, what is the specific relevance to this organization?

#SourceDateTLPAdmiralty SourceAdmiralty InfoKey ClaimRelevance to CelltronX
TRG-001Peer company incident — MobileTech IL Ltd.3 weeks agoAMBERB (usually reliable)2 (probably true)Iranian-nexus actor compromised Israeli telecom via contractor maintenance account — same vendor type as CelltronX usesCritical — CelltronX uses NetSys Solutions Ltd., same access model as confirmed victim
TRG-002CERT-IL TLP:AMBER bulletin10 days agoAMBERB (usually reliable)2 (probably true)Iranian-nexus actor targeting Israeli telecom billing and roaming data via vendor API abuseHigh — CelltronX billing system API accessible to contractors; same target data type
TRG-003Domain intelligence (OSINT)8 days agoWHITEC (fairly reliable)3 (possibly true)Three domains registered impersonating CelltronX targeting NOC staff logins and fake software update luresHigh — Pre-attack infrastructure; phishing campaign against NOC engineers imminent
TRG-004Threat actor profile — CEDAR-SIGNALInternalA (completely reliable, internal analysis)1 (confirmed)CEDAR-SIGNAL cluster: AiTM credential theft, telecom targeting, disruption capability, active against Israeli sector since Q3 2024High — CEDAR-SIGNAL capability matches all three triggers

Synthesized threat picture:

An adversary assessed as Iranian-nexus is currently operating against Israeli telecommunications providers through contractor supply chain access. TRG-001 confirms they successfully used this vector against a peer organization. TRG-002 confirms billing and roaming data is their target. TRG-003 shows they have already built phishing infrastructure specifically targeting CelltronX's NOC staff. TRG-004 provides the actor profile.

CelltronX must treat its contractor access as potentially compromised until audited. This is not a theoretical threat — the adversary has already executed this attack against a peer company in the same sector using the same access model.

git add 01-trigger-intelligence/
git commit -m "ASSESS-002: trigger intelligence — 4 signals, synthesized picture: CEDAR-SIGNAL active, contractor access at risk"

Step P4: Crown Jewels Analysis — What Would Be Irreversible

The crown jewels analysis answers: if this threat actor gets in, what can they reach, and what would be the consequence? It forces scope discipline — the detection backlog only covers paths to crown jewels.

IDAssetWhy CriticalBreach ConsequenceCurrent ExposurePriority
CJ-001Subscriber billing records (4.2M customers)Revenue engine; INCD-CID designated Critical InfrastructureMass identity theft; Israeli PPL fine; TASE disclosure obligation; reputationalAPI accessible to contractors; no rate limiting; logs not in SIEMCritical
CJ-002Network Management System / ENMControls routing for all 4.2M subscribers; PM Office contract dependencyService disruption or selective blackout; national security incidentENM CVE-2023-44481 unpatched (v21.4.8); contractor remote access activeCritical
CJ-003PM Office connectivity contract ($210M)Government CII designation; politically sensitiveContract loss; national security incident; government reviewDependent on CJ-002 protection; no independent access controlCritical
CJ-004SS7 signaling networkSubscriber call/SMS/location interception capabilitySIGINT value to foreign intelligence; subscriber surveillanceAccessible via NMS once NMS is compromised; not logged or monitoredCritical
CJ-005Roaming data (UAE, Cyprus, Romania)International travel patterns; foreign intelligence valueForeign intelligence collection on Israeli subscribersAPI access from NMS segment; no anomaly detectionHigh
CJ-006NOC contractor access credentialsSingle gateway to CJ-001 through CJ-005All above compromised via one actorAiTM phishing infrastructure targeting NOC already deployed (TRG-003)Critical

The exposure scoring reveals the cascade. CJ-006 (contractor credentials) scores highest exposure because phishing infrastructure targeting it already exists. But CJ-002 (ENM) scores highest independently because a known CVE is unpatched and exploitable without any credential theft at all. An attacker who does not have valid credentials can still enter via CVE-2023-44481. These are two independent paths to the same crown jewel.

git add 02-crown-jewels/
git commit -m "ASSESS-002: crown jewels — 6 assets; CJ-006 (contractor creds) and CJ-002 (ENM) highest exposure; CVE-2023-44481 unpatched"

Step P5: Attack Scenarios — From Crown Jewels to Kill Chain

Each scenario maps a realistic path from the threat actor's current capability (documented in triggers) to a crown jewel, with every step mapped to ATT&CK and coverage assessed at each step.

Scenario 1 (HIGH risk): NOC Contractor Compromise → ENM Access → SS7 Collection

Why this scenario is rated HIGH: Same vendor access model as TRG-001 victim. Same AiTM technique documented in TRG-002. CEDAR-SIGNAL is documented using this exact path. CelltronX uses the same contractor type. Likelihood is not "could happen" — it is "this has already happened to our peer."

Kill chain:

StepTechniqueProcedureCurrent Coverage
1T1566.001Spearphishing with ISO lure → fake NOC tool update notification (TRG-003 infrastructure)Partial — no ISO-specific rule; ATP SCL threshold not tuned
2T1557AiTM proxy intercepts NOC employee M365 session during fake authentication pageNone — no AiTM session token detection
3T1078.002Replays stolen credentials to contractor VPN — no MFA re-challenge on valid sessionNone — no anomalous contractor VPN auth rule
4T1190Exploits CVE-2023-44481 on unpatched ENM v21.4.8 (alternative to step 3 — no credentials needed)None — no WAF rule for ENM; patch not applied
5T1021.001RDP from contractor DMZ to ENM/NMS operational segmentPartial — rule exists but 30-min alert delay
6T1040SS7 MAP queries from NMS access to subscriber recordsNone — SS7 MAP traffic not logged
7T1048.003Exfiltration via DNS TXT record tunnelingNone — DNS TXT responses not monitored

Coverage verdict: 0 of 7 techniques have adequate coverage. Steps 3, 4, 6, and 7 have complete absence.

The CVE finding is the most urgent issue. Step 4 (CVE-2023-44481) means the adversary does not need to succeed at steps 1–3. If the phishing fails, they can still enter directly. Patching the ENM is not a detection task — it is an emergency remediation item that must run in parallel with detection development.

Scenario 2 (HIGH risk): Billing API Credential Abuse → Subscriber Record Exfiltration

Kill chain:

StepTechniqueProcedureCurrent Coverage
1T1078.003Stolen billing API credentials (obtained from compromised contractor laptop or code repository)None — billing API logs not in SIEM
2T1530Authenticated bulk API calls mimicking legitimate maintenance queries; below any volume thresholdNone — no rate limiting; no volume anomaly rule
3T1048Exfiltration over HTTPS using valid API sessionNone — indistinguishable from legitimate API traffic without volume baseline

Coverage verdict: 0 of 3 steps covered. Primary blocker: billing API logs are not in the SIEM.

Unlike Scenario 1, this attack leaves no anomalous process execution, no lateral movement, no registry changes. It looks identical to legitimate contractor maintenance queries until a volume baseline is established. The only detection surface is API log anomaly detection — and those logs are not in the SIEM.

Scenario 3 (MEDIUM risk): NMS Configuration Change → PM Office Connectivity Disruption

Prerequisite: Requires successful Scenario 1 first. Current likelihood: Low (actor appears in reconnaissance/pre-positioning phase). Impact: Critical.

This scenario is rated MEDIUM not because the impact is manageable — a targeted disruption of government connectivity is a national security incident — but because it requires Scenario 1 as a prerequisite. If Scenario 1 detections are deployed, Scenario 3 is automatically harder to execute.

git add 03-threat-model/
git commit -m "ASSESS-002: scenarios — SCN-001/002/003; 0/7 techniques covered in SCN-001; CVE-2023-44481 emergency patch required"

Step P6: Detection Backlog — Intelligence Translated to Sprint Work

The detection backlog is not a wishlist. Every item has: which scenario it covers, what pre-condition must be true before the rule can be written, and which sprint it belongs to given the team's 6–8 rule capacity.

IDDetectionScenarioATT&CKPre-conditionPrioritySprintBlocked?
DET-001Contractor VPN auth from non-corporate ASN, off-hoursSCN-001 step 3T1133 + T1557ASN enrichment on VPN logsP1Sprint 1No
DET-002RDP from Contractor DMZ to ENM segment, off-hoursSCN-001 step 5T1021.001NoneP1Sprint 1No
DET-003ENM authentication failure spike (CVE-2023-44481 probing)SCN-001 step 4T1190ENM logs ingestedP1Sprint 1No
DET-004NetSys contractor account authentication outside business hoursSCN-001 step 3T1078.002Contractor account list from HRP1Sprint 1No
DET-005LSASS dump via comsvcs.dll on ENM hostsSCN-001T1003.001Sysmon on ENM hostsP1Sprint 1Check — Sysmon deployment
DET-006Billing API bulk query from non-standard IPSCN-002 step 1T1078.003Billing API logs in SIEMP1Sprint 2Yes — no SIEM integration
DET-007Billing API call volume anomaly (>N records in M minutes)SCN-002 step 2T1530Billing API logs in SIEM + 30-day baselineP1Sprint 3Yes — baseline needed
DET-008NMS configuration write outside change windowSCN-003T1565.001NMS audit logs in SIEMP1Sprint 2Yes — NMS logs not in SIEM

Blocked items are explicit. DET-006, DET-007, and DET-008 cannot be written yet because the required log sources are not in the SIEM. The backlog separates these clearly so the sprint plan is realistic: the detection team spends Sprint 1 deploying DET-001 through DET-005 while the engineering team works on the billing API and NMS log pipelines for Sprint 2.

Compensating controls for blocked items:

  • DET-006 (blocked): SOC runs daily manual review of billing API access logs — 30-minute review of any access volumes > 10,000 records or from non-approved IPs
  • DET-008 (blocked): Network Operations conducts weekly NMS configuration diff against last-known-good snapshot

These are not detection rules. They are temporary, human-executed compensating controls that are explicitly documented as such — with an owner and a sunset date (when the log pipeline is complete).

git add 04-detection-backlog/
git commit -m "ASSESS-002: detection backlog — 8 items; Sprint 1: DET-001–005 (unblocked); Sprint 2: DET-006+008 unblock work; compensating controls documented"

The 72-Hour Immediate Action Plan

Not everything waits for the sprint cycle. Some actions are so high-confidence and low-effort that they must happen this week, before the detection backlog is built.

ActionTriggerOwnerDeadlineRisk of Not Acting
Block 3 impersonation domains at DNS/email gatewayTRG-003 — phishing infrastructure targeting NOC staff is active nowSOC24 hoursNOC engineer clicks a link this week before detections are deployed
Alert NOC team about the threatTRG-003 — specific, current targetingCTI + NOC Lead24 hoursSocial engineering succeeds because staff don't know they're being targeted
Request emergency ENM patch (CVE-2023-44481) from EricssonENM CVE is directly exploitable; Scenario 1 step 4 bypasses all credential detectionNOC Engineering + Ericsson TAC48 hoursAdversary exploits CVE before detection is in place — SCN-001 executes without step 1–3
Audit all active NetSys contractor sessions in past 30 days against known working hoursTRG-001 — same vendor confirmed compromised at peer companySOC + CISO24 hoursOngoing compromise is already present; delay extends dwell time
Force MFA re-enrollment for all NetSys contractor accountsTRG-002 — AiTM session token theft bypasses existing MFA; re-enrollment forces fresh credential bindingIT + CISO decision48 hoursStolen session tokens remain valid for extended periods after AiTM interception
Send formal security audit request to NetSysTRG-001 — peer company victim used same vendorCISO → NetSys CISO (contract manager)24 hoursIf NetSys is compromised, CelltronX access is actively at risk

The 72-hour plan is the bridge between intelligence and action — the things the analysis team recommends that do not require sprint cycles, CAB approvals, or detection engineering capacity. They are immediate, high-signal, low-cost. The CISO needs to make one decision: suspend NetSys remote access pending the audit (Option A — disrupts maintenance) or maintain access with enhanced monitoring and force MFA re-enrollment (Option B — lower disruption, requires DET-001 and DET-004 live within 24 hours).


Deliverables — The CISO Decision Brief

The executive brief is one page. It leads with the bottom line, quantifies the exposure, and ends with a specific decision request.


Executive Brief — CEDAR-SIGNAL Threat to CelltronX TLP:AMBER | For: CISO Dr. Katz | Date: 2024-10-15

Bottom line: An Iranian-linked threat group that recently compromised a peer Israeli telecom through a contractor access arrangement is targeting CelltronX through the same vector. Three impersonation domains targeting our NOC staff were registered this week. A critical vulnerability in our network management platform is unpatched and directly exploitable.

Why now: The same hacker group that compromised MobileTech IL's network management system three weeks ago is now deploying phishing infrastructure against CelltronX. CelltronX uses the same type of contractor arrangement as the victim. The attacker has already built fake login portals that mimic our internal systems.

Current exposure:

What's at RiskExposureWhy
ENM (network management)CriticalCVE-2023-44481 unpatched — accessible without credentials
Contractor VPN accessHighSame access model as confirmed victim; phishing active
Billing records (4.2M subscribers)HighBilling API logs not monitored; no rate limiting
PM Office connectivity contractHighDependent on ENM protection; no independent failsafe

What we are doing this week: Block 3 impersonation domains. Brief NOC team. Request emergency Ericsson patch. Audit NetSys contractor access history.

Decision required — by end of week: Suspend NetSys remote access for 5 business days pending security audit (disrupts scheduled maintenance) or maintain access with enhanced monitoring and mandatory MFA re-enrollment (requires two detection rules live within 24 hours).

Recommendation: Option B (maintain access, force MFA, deploy rules within 24h) escalating to Option A if NetSys does not provide clean audit confirmation within 5 business days.

Confidence: High — based on 4 independent triggers including a government advisory and a confirmed peer incident.


The Git History: What a Proactive Assessment Looks Like

4d7a2f1 ASSESS-002: deliverables — CISO brief, technical brief, SOC 72h plan
2c9b5e8 ASSESS-002: detection backlog — 8 items, Sprint 1 unblocked, compensating controls
7f3a1c4 ASSESS-002: scenarios — SCN-001 (0/7 covered), SCN-002 (0/3 covered), CVE emergency patch
1e6d8b3 ASSESS-002: crown jewels — 6 assets, CJ-006 highest exposure, ENM CVE critical
9a2c5f7 ASSESS-002: trigger intelligence — 4 signals; synthesized: contractor access at risk
6b8f4d1 ASSESS-002: scope — signed off; 3 PIRs, 6-8 sprint capacity, 6mo INCD deadline
3e1a7c9 ASSESS-002: intake — 4 triggers, INCD audit in 6mo, engineering capacity documented
0d4b6f2 ASSESS-002: scaffold initialized

Every commit in a proactive assessment is a decision about what the threat means for this organization. The git history is how the CISO can see that the assessment was systematic — not that someone read a threat report and wrote a list of controls.


Key Lessons

Intelligence without context is noise. The CERT-IL bulletin (TRG-002) on its own says "Iranian actor targeting Israeli telecom." That describes half the sector. What makes it actionable is TRG-001 (confirmed peer incident using the same contractor), TRG-003 (impersonation domains targeting CelltronX specifically), and TRG-004 (actor profile matched to the other triggers). Each trigger alone is incomplete. Together they form a specific, actionable threat picture.

The crown jewels define scope. Without the crown jewels analysis, the detection backlog would be generic best-practice rules. With it, every backlog item traces to a specific asset (ENM, billing system, PM Office contract) and a specific attack path. The SOC engineer writing DET-001 knows which crown jewel they are protecting and which scenario it covers. That context changes how the rule is tuned.

Blocked items must be visible. The most dangerous outcome of a detection backlog is reporting "12 rules delivered" when 4 of them cannot be written because the log sources are not ingested. The sprint plan must separate "rule ready to deploy" from "rule blocked on pipeline." Blocked items need compensating controls and owners, not placeholders.

A CVE can render all credential detection moot. ENM CVE-2023-44481 does not need the contractor's credentials to be stolen. It is a direct exploitation path. Writing a perfect AiTM detection rule (DET-001) has zero impact on an adversary who uses the CVE instead. The technical brief to the CISO must make this explicit — detection investment and patch management must be coordinated, not sequential.

The 72-hour plan is not an afterthought. The most immediate risk (NOC staff clicking phishing links from TRG-003 infrastructure before any detection is deployed) is addressed by the 72-hour plan, not the sprint cycle. Domain blocking, user awareness, and contractor access audit cost almost nothing and can happen today. They buy the time needed to build the real detection coverage.


This scenario is training assignment A02 from the CTI as a Code repository. The full evidence set, template, and worked solution are available there.

Tags: Threat Intelligence · CTI · Proactive Assessment · Threat Modeling · Detection Engineering · MITRE ATT&CK · Blue Team · Supply Chain Security