CSRP — Cyber Security Responder & Practitioner
Lesson 25: Cyber Kill Chain and Attacker Mindset
Textbook
Cyber School · BDO Ziv Haft
CSRP LESSON 25 — TEXTBOOK
Module 4 — Offensive Cyber & Vulnerability Analysis · Lesson 1 of 8 · Course Lesson 25 of 40 — M4 OPENS Title: Cyber Kill Chain and Attacker Mindset
Deep-reading companion. ~6,500 words.
0 · Why this lesson, and the M4 arc
Module 4 is the pivot of CSRP. For 24 lessons you’ve been a defender — building the firewall (L9), the IDS (L12), IAM (L13), MFA (L14), DLP (L15), EDR (L16), mail security (L17), web proxy (L18), regulation (L19), cloud (L20), and across Module 3 the full AI-security stack (L21-L24). You know how to construct defenses. You know how to verify them. You are a Cyber Practitioner.
From today through L32, you study the attacker. Not to become one. To understand the attacker’s logic, sequence, and trade-offs well enough that your defense improves. Threat hunters, incident responders, red teamers, and even ordinary SOC analysts all share this fluency. Without it, you respond to events one at a time and miss the chain. With it, you see the chain and interrupt it early.
L25 introduces the canonical structured model — the Lockheed Martin Cyber Kill Chain, published in 2011. It’s the shared vocabulary of every security operations center, every threat-intel report, every red-team debrief, every regulatory incident report. It is not the only model — MITRE ATT&CK provides finer-grained technique IDs (we’ll layer those in L26 onward), and the Unified Kill Chain integrates both — but the original Lockheed model is what you learn first because it teaches the right defender mental model: attacks are sequences, and defenders can break the chain at ANY phase to stop the attack.
L25 also opens the M4 ethical framing. Every offensive technique we discuss across M4 is paired with the legal boundaries: authorized targets only, lab environments only, written rules of engagement for any “real” pen-testing role you may later take. Using the techniques outside that boundary is illegal in every jurisdiction where CSRP operates. If in doubt — ask before you act.
By the end of L25 you can: recite the 7 phases with defender leverage points; read multi-source logs and map events to phases; build the preventive / detective / response control matrix; design attack paths through layered enterprise networks; recommend segmentation controls that break the path; apply the “break the chain anywhere” principle to defensive prioritization.
1 · The Lockheed Martin Cyber Kill Chain — 7 phases
The Kill Chain originated in military terminology — the sequence of actions required to destroy a target. Lockheed Martin (specifically researchers Eric Hutchins, Michael Cloppert, and Rohan Amin) adapted it to cyber-attack analysis in their 2011 paper “Intelligence-Driven Computer Network Defense Informed by Analysis of Adversary Campaigns and Intrusion Kill Chains.” The model presents 7 phases an attacker traverses from initial interest in a target through to objective completion.
1.1 Phase 1 — Reconnaissance
The attacker gathers information about the target. This phase happens entirely outside the defender’s perimeter, often using public sources and entirely legal information-gathering — at least until the attacker starts probing systems directly.
What attackers do here: scrape company websites; index LinkedIn employees and their job histories; collect job postings (which reveal the tech stack: “we’re hiring a Cisco ASA admin” tells the attacker you run Cisco ASA); enumerate DNS records (mail servers, VPN gateways, subdomains); look up Whois data; examine GitHub for accidentally-published configs or credentials; collect sample emails to learn the company’s email format; monitor social media for employee complaints or news; identify executives (high-value phishing targets).
What the defender can do here: surprisingly limited. The information is public; making it less so is hard. The leverage points are: – Reduce attack surface at the source: don’t publish unnecessary configs to GitHub; don’t reveal too much in job postings; train employees on social-engineering awareness regarding LinkedIn requests. – Honey-tokens: plant fake LinkedIn profiles, fake email addresses on the company website, fake employee accounts — any access attempt against these is a strong attacker signal. – Canary domains: register decoy subdomains. Any DNS query for them suggests targeted reconnaissance. – Threat-intel feeds: some recon activities (port scans from known-bad IPs, querying recently-registered look-alike domains) are visible via intel feeds.
The defender’s strategic answer to Recon: accept that some attacker knowledge is unavoidable, and DESIGN ASSUMING IT. Your defenses must work even when the attacker knows your stack, your employee names, and your email format — because they will.
1.2 Phase 2 — Weaponization
The attacker combines an exploit (something that triggers a vulnerability) with a payload (something that runs once the vulnerability is triggered) to produce a deliverable artifact. This phase happens entirely on attacker infrastructure.
Examples: a crafted PDF that exploits a known PDF reader vulnerability and drops a backdoor; an Office document with a macro that downloads a stage-2 payload; a malicious USB stick with autorun; a watering-hole website that exploits browser vulnerabilities and serves browser-resident malware; an LLM-prompt-injection campaign embedded in apparently-innocuous documents.
What the defender can do here: essentially nothing direct. The attacker controls their own environment. The leverage is INDIRECT: threat-intel feeds may identify the exploit/payload artifact’s hash, signature, or C2 infrastructure before delivery. Patching known vulnerabilities reduces what’s exploitable in your environment. Application hardening reduces what payloads can do once they run.
1.3 Phase 3 — Delivery
The attacker transmits the weaponized artifact to the target environment. This is the first phase the defender has direct visibility into — and one of the highest-leverage defensive opportunities.
Common delivery vectors: phishing emails (with attachments or links); watering-hole websites (the attacker compromises a site the target visits); USB drops; supply-chain compromise (the attacker plants malware in software the target installs); social engineering (the attacker calls and tricks an employee into deploying the artifact); cloud-misconfig exploitation (publicly-exposed S3, Azure Blob).
Defender leverage at Delivery is HIGH: – Mail security (L17): SPF/DKIM/DMARC enforcement, attachment sandboxing, link inspection, AI-augmented content filtering. Mail-borne delivery still accounts for ~70% of incidents in the IBM Cost of Data Breach reports. – Web proxy (L18): URL filtering, reputation, content inspection, malicious-domain blocking. Catches watering-hole and link-click attacks. – USB controls: device control via EDR (L16), Group Policy, OS-level enforcement. – Supply-chain hardening: SBOM (Software Bill of Materials) reviews, vendor security questionnaires, code-signing enforcement. – User awareness training: ongoing, simulated phishing.
The defender’s strategic answer to Delivery: layered controls. No single layer catches everything. Mail + Proxy + EDR + Training + USB-control combine for high coverage.
1.4 Phase 4 — Exploitation
The delivered artifact triggers and exploits a vulnerability. Code now RUNS in the target environment for the first time, with the privileges of whatever user/process triggered it.
Examples: a user opens the malicious PDF and the PDF reader is exploited; a macro in an Office doc fires and spawns PowerShell; a browser exploit triggers when the user visits the watering-hole site; an exposed RDP service is exploited via a known CVE; a public-facing web app is exploited via SQL injection or remote code execution; a cloud workload is exploited via a publicly-exposed metadata-service vulnerability.
Defender leverage at Exploitation is HIGH: – Patching: the most under-rated defense in cybersecurity. If the vulnerability isn’t there, the exploit fails. – Endpoint hardening: ASLR, DEP, EMET, control-flow guard. These don’t eliminate vulnerabilities but raise the cost of exploitation. – Application allow-listing: the user can’t run a process that isn’t on the allow-list. Stops macro-spawned PowerShell cold. – EDR exploit detection (L16): behavioral signatures for shellcode patterns, process-creation chains (Outlook → PowerShell), and known exploitation techniques. – Browser hardening: site-isolation, sandboxing, blocked downloads from non-trusted sources.
The defender’s strategic answer to Exploitation: patch fast, harden everywhere, deploy EDR with behavioral detection enabled. This phase is where competent defenders break ~30-40% of attacks that get past Delivery.
1.5 Phase 5 — Installation
The exploited target now has attacker-controlled code running, but only IN-MEMORY. The attacker installs persistence — a mechanism to maintain access even after a reboot or process kill.
Common persistence mechanisms: scheduled tasks (Windows); cron jobs (Linux); Windows services; registry Run keys; WMI event subscriptions; backdoored startup folders; rootkits (kernel-level); webshells on web servers; remote-access tool (RAT) installation; legitimate-tool abuse (TeamViewer / AnyDesk planted by attacker); Active Directory persistence (golden ticket, silver ticket, DCSync).
Defender leverage at Installation is HIGH: – EDR persistence detection (L16): all major EDR products detect common persistence techniques out of the box. – Sysmon + log forwarding: with the right rule set, Sysmon catches scheduled-task creation, service installation, registry-Run-key modification, and WMI event subscriptions. – File integrity monitoring: catches changes to critical system files. – Active Directory tier-0 protection: protect domain admins; require privileged-access workstations; monitor DC events. – Webshell detection: signature-based + behavioral.
The defender’s strategic answer to Installation: catch persistence FAST. The longer attacker persistence remains, the deeper they entrench. Industry data shows attackers establish persistence within 30 minutes of exploitation in most cases.
1.6 Phase 6 — Command & Control (C2)
The attacker establishes an outbound communication channel from the compromised target to attacker-controlled infrastructure. This enables remote control, command execution, and stage-2 payload delivery.
Common C2 channels: HTTPS to attacker-controlled domains; DNS tunneling; ICMP tunneling; Slack/Discord/Telegram bot APIs; legitimate cloud services (Pastebin, GitHub, Dropbox) used as C2; encrypted custom protocols on non-standard ports.
Modern attackers prefer C2 that BLENDS with legitimate traffic: HTTPS to a CDN-fronted attacker domain looks identical (at first glance) to normal web browsing. DNS-tunnel C2 hides in normal DNS traffic. Slack-API C2 looks like normal Slack usage.
Defender leverage at C2 is HIGH: – Egress filtering (L9): default-deny outbound from servers; allow-list only the destinations they need. Stops most C2 cold. – Web proxy (L18): TLS-inspecting proxy can decrypt + inspect HTTPS C2 (with appropriate legal/privacy framing). – DNS monitoring: alert on DNS queries to unusually-recently-registered domains, suspicious TLDs, or non-resolvable queries (DNS-tunnel signal). – IDS/IPS (L12): catches known C2 protocols, JA3 fingerprints of common offensive tools (Metasploit, Cobalt Strike). – Threat-intel domain/IP blocks: feed-based blocking of known C2 infrastructure.
The defender’s strategic answer to C2: assume the attacker WILL try multiple channels. Default-deny egress + DNS monitoring + threat-intel feeds + IDS combine to catch most.
1.7 Phase 7 — Actions on Objectives
The attacker completes the goal. By this point they have persistence, remote control, and lateral movement. What they DO depends on their motivation: – Data theft (espionage / IP): exfiltrate trade secrets, customer data, M&A data. – Financial (BEC + extortion): redirect wires, lock systems for ransom, extort customers with stolen data. – Ransomware: encrypt files, demand payment. – Destruction: wipe systems (NotPetya-style). – Disruption: DDoS internally; disable production.
Defender leverage at Actions is MEDIUM (not high — it’s late): – Data Loss Prevention (DLP) (L15): catches outbound exfiltration of classified data. – UEBA (L22): catches unusual data-access volumes, off-hours activity, abnormal admin behavior. – Network segmentation (L9 / L25.3): limits how far the attacker can pivot from initial foothold. – Backup integrity (offline / immutable): the ransomware-survival control. – Incident response readiness: by Phase 7 you’re in IR mode — readiness determines outcome.
The defender’s strategic answer to Actions: prevention failed; minimize blast radius. Segmentation, DLP, UEBA, backups, IR readiness.
2 · How modern attacks diverge from the linear model
The Lockheed model is sequential and conceptual. Real 2026 attacks often:
Parallelize: an attacker may execute multiple Kill Chains simultaneously against different targets, then converge on the highest-success path.
Skip phases: a “living off the land” attacker who lands via stolen credentials skips Weaponization, Delivery, and Exploitation entirely — they have valid credentials, they just log in.
Loop: persistence (Installation) followed by additional Recon (now from inside), then more Lateral Movement (which Lockheed doesn’t call out separately — MITRE ATT&CK does), then Actions.
Chain Kill Chains: a successful Actions-on-Objectives in one organization (stealing partner credentials from Org A) becomes the Reconnaissance asset for the next attack (against Org A’s partner Org B). Supply-chain attacks like SolarWinds/3CX exemplify this.
The lesson: the Kill Chain is a TEACHING MODEL. Use it as the structure for thinking about attacks. Don’t expect every real attack to fit cleanly. The “break the chain” insight holds even when the chain is fuzzy.
3 · MITRE ATT&CK — the granular partner framework
MITRE ATT&CK (Adversarial Tactics, Techniques, and Common Knowledge) is the modern, finer-grained framework. While Kill Chain has 7 phases, ATT&CK has 14 tactics (the higher-level goals — reconnaissance, initial access, execution, persistence, privilege escalation, defense evasion, credential access, discovery, lateral movement, collection, command and control, exfiltration, impact, plus a recently-added “resource development”) and hundreds of techniques within each tactic.
ATT&CK is the working analyst’s daily reference. Every threat-intel report cites ATT&CK technique IDs. Every EDR product maps its detection rules to ATT&CK. CSRP labs from L26 onward cite ATT&CK IDs alongside Kill Chain phases.
The two frameworks complement: Kill Chain teaches lifecycle and “break the chain”; ATT&CK teaches the techniques within each phase.
4 · The attacker mindset — three constraints
Attackers are not magic. They are economic actors optimizing across three constraints:
Time. Every step costs effort. Recon takes hours to weeks. Tool development takes days. Initial access campaigns take weeks. Lateral movement takes hours per pivot. Patient nation-state actors can dwell for months — but they’re the exception.
Detection cost. Some techniques are loud (Mimikatz dump fires every EDR product immediately). Some are quiet (an LSASS dump via the Comsvcs.dll MiniDump method evades many detections). Quiet costs more time and skill.
Goal payoff. The attacker’s mission determines acceptable trade-offs. Quick credential theft for fraud accepts higher detection risk for speed. Espionage demands stealth above all. Ransomware demands speed AND broad reach.
The defender’s strategic insight: make every attacker step expensive in time OR detection risk OR both. A defender who raises both costs forces the attacker to either invest more (which only well-funded adversaries will do) or pick an easier target (the low-hanging-fruit principle).
5 · The “living off the land” preference
Modern attackers prefer to use BUILT-IN TOOLS rather than custom malware: – Windows: PowerShell, WMI, PsExec, certutil, bitsadmin, scheduled tasks, BITS jobs, RDP, registry edits. – Linux: ssh, scp, cron, systemd services, named pipes, common system tools. – Both: legitimate remote-access tools (AnyDesk, TeamViewer) installed for legitimate IT use, then abused.
Why: AV and signature-based EDR catch CUSTOM malware. PowerShell is in every Windows install. RDP is everywhere. Distinguishing “legitimate admin PowerShell” from “attacker PowerShell” requires behavioral analysis, not signatures.
The defender’s strategic answer to LotL: behavioral detection over signatures. UEBA (L22) baselines per-user activity. EDR with behavioral rules (L16) detects suspicious process trees. Audit logging captures the artifacts (PowerShell -enc, suspicious WMI calls). MITRE ATT&CK’s “LOLBins” project catalogs the abusable binaries.
6 · The network-zone path
Most enterprises segment into 4-6 trust zones. The attacker’s job is to cross zone boundaries. The defender’s job is to make crossings expensive.
Typical zones: – Internet — outside, no trust. – DMZ — internet-facing services, partial trust. Web, mail, VPN gateway, sometimes reverse proxies. – Internal LAN — user workstations, moderate trust. – Admin network — privileged-access workstations (PAW) and jump hosts, high trust. – Server network — application servers and databases, high trust. – (sometimes) OT/SCADA — industrial control, critical trust.
A typical attacker path: Internet → DMZ pivot (exploit a web service) → Internal LAN (phishing-landed user) → Admin network (stolen admin credentials) → Server network (admin-RDP to a database server) → exfil out. Each crossing is a defensive opportunity.
Crossings to defend: – Internet → DMZ: perimeter firewall (L9), WAF, DDoS protection, IDS (L12). – DMZ → Internal LAN: internal firewall, no direct routing, jump-host required, IDS at the boundary. – Internal LAN → Server: tight ACLs, no direct routing, jump-host required. – Internal LAN → Admin: MFA enforced (L14), restricted RDP, audit logging. – Admin → Server: MFA + privileged-access workstations + UEBA (L22). – Server → Internet (egress): default-deny + DLP (L15) + egress proxy.
Lab 25.3 puts you into a Packet Tracer environment with the 5 zones and intentional weaknesses. Find the weaknesses; recommend the fixes.
7 · “Break the Chain” — the defender’s operating principle
You don’t have to stop EVERY phase. You only have to break the chain at ONE phase to stop the attack.
This is the defender’s hope in a world of overwhelming attacker capability. The Kill Chain is (mostly) sequential. Break it anywhere and the attacker’s objective fails.
The strategic question becomes: WHERE to break.
Earliest is best. Pre-Delivery has lowest blast radius. The attacker hasn’t reached your environment yet. Hardening Recon (limited as it is) and disrupting Delivery wins more than playing Whac-A-Mole in C2.
Cheapest is best per dollar. Well-known defenses — MFA, patching, EDR, mail security — beat exotic products. The exotic XDR-AI-magic-quadrant tool may help, but if you haven’t deployed MFA universally, you’re spending in the wrong order.
Highest-detection is best per analyst-hour. Controls that produce clear, actionable alerts let your SOC team focus. Controls that just produce noise create alert-fatigue and degrade your real detections.
The Pareto reality
Industry data consistently shows ~80% of real-world attacks break on six controls:
- Phishing-resistant MFA (L14) — defeats the credential-theft → re-use → privilege-escalation chain that underlies most attacks.
- Endpoint patching (Windows + Linux) — removes the vulnerabilities most exploits target.
- Mail security (L17) — catches phishing Delivery at the gate.
- EDR with current signatures + behavioral rules (L16) — catches Exploitation and Installation.
- Egress filtering (L9) — catches C2 establishment.
- Segmentation between user and server networks (L9 / L25.3) — limits Actions blast radius.
An organization with these six fully deployed is harder to compromise than ~90% of mid-sized enterprises. Module 2 trained you to deploy them; Module 4 helps you UNDERSTAND why each works.
8 · The role transition
After Module 3 you’re a Cyber Practitioner — you can BUILD the defenses. After Module 4 you’ll be a Cyber Analyst — you can THINK like the attacker, INVESTIGATE incidents, and CORRELATE events across the Kill Chain.
The analyst role demands: – Reading multi-source logs and correlating timestamps. – Mapping events to a framework (Kill Chain + ATT&CK). – Recognizing attack patterns and the chain. – Articulating the gap that allowed the attack. – Recommending controls that close the gap.
L25 is the first day of that role. The labs today (mapping a real attack timeline, building the control matrix, designing the path) ARE the analyst’s daily work — just on prepared data. M5 (L33-L40) will put it on live data with a SIEM. Today builds the framework.
9 · The labs in detail
Lab 25.1 — Map an Attack to the Cyber Kill Chain. You read ~25 events across multiple log sources covering a 36-hour incident at Meridian. Tag every event with a Kill Chain phase. Identify which phases the attack used vs. skipped. Write a 200-word analyst summary covering initial vector, persistence, exfiltration. The lesson: correlating log sources around a timeline is the analyst’s daily work.
Lab 25.2 — Defensive Control Mapping. Given Lab 25.1’s mapped attack, populate the 7×3 control matrix (Preventive / Detective / Response per phase). Mark each cell [have] or [missing] for Meridian. Prioritize Top 5 improvements with rationale + effort estimate. Write a 1-page roadmap presentation outline. The lesson: offensive thinking IS defensive planning; structured frameworks support structured prioritization.
Lab 25.3 — Attack Path Design in Cisco Packet Tracer. Open the pre-built 5-zone topology. Identify zone boundaries. Mark exposed services. Find segmentation weaknesses. Draw attacker movement path Internet → Server. Recommend specific firewall / ACL / MFA / IDS controls for each crossing. The lesson: attack-path visualization is a senior-analyst skill; the same diagram both informs the attacker and the defender.
Challenge — Break the Chain. Browser-based, 25 minutes. Map 30 events to phases. Identify the earliest break-point where Meridian’s existing controls catch the chain. Identify ONE missing control that breaks the chain earlier. Submit ULC{…} after validator confirms. The lesson: this is the structure of every incident retrospective you’ll write.
10 · Connections to CSRP modules
- M2 L9-L20 — every defensive control gets re-contextualized today within the Kill Chain matrix. The L17 mail-security work? That’s Delivery defense. The L16 EDR work? Exploitation + Installation. The L14 MFA work? Defeats credential-reuse during lateral movement.
- L21-L24 M3 — the verification discipline + HITL pattern transfer to incident analysis. Don’t trust the SIEM’s auto-classification; verify the chain.
- M4 L26-L32 — each lesson zooms into one Kill Chain phase: L26 = Recon detail; L27 = scanning (pre-Delivery); L28 = vuln assessment (Exploitation preparation); L29 = Exploitation hands-on; L30 = Web App Exploitation; L31 = Social-engineering Delivery; L32 = Credential-attack techniques used during Installation / Lateral.
- M5 L33-L40 — Kill Chain + ATT&CK become daily SOC vocabulary; capstones (L38/L39) integrate the full cycle.
11 · Key terms (preview of file 13)
- Cyber Kill Chain — Lockheed Martin’s 7-phase attack-lifecycle model.
- Reconnaissance / Weaponization / Delivery / Exploitation / Installation / Command & Control / Actions on Objectives — the 7 phases.
- MITRE ATT&CK — the granular technique-level framework.
- Living off the Land (LotL) — using built-in OS tools instead of custom malware.
- LOLBin — Living Off the Land Binary; a legitimate OS binary abusable by attackers.
- Preventive / Detective / Response controls — the 3 control types per Kill Chain phase.
- Network zone / segmentation — trust-based network partitioning.
- Attack path — the sequence of compromises from initial access to objective.
- Break the chain — the defender’s principle of disrupting any single phase to stop the attack.
- Pareto controls — the 6 controls that block 80% of real attacks.
12 · Further reading
- Hutchins, Cloppert, Amin — “Intelligence-Driven Computer Network Defense Informed by Analysis of Adversary Campaigns and Intrusion Kill Chains” (Lockheed Martin, 2011)
- MITRE ATT&CK — attack.mitre.org
- Verizon Data Breach Investigations Report (DBIR) — annual; cites real-world Kill Chain phase distribution
- IBM Cost of a Data Breach Report — annual
- Mandiant M-Trends — annual; modern dwell-time and Kill-Chain analysis
13 · Self-test
- Recite the 7 Kill Chain phases in order.
- For each phase, name one control type (preventive / detective / response).
- Define “Living off the Land” and explain its defensive implication.
- State the “break the chain” principle.
- Name the 6 Pareto controls.
- Distinguish Exploitation from Installation by one sentence.
- Why does the Kill Chain matter for SOC vocabulary even when modern attacks diverge from it?
14 · Closing
You’ve covered the framework. Tomorrow (L26) you go into Phase 1 in detail — Open-Source Intelligence and Reconnaissance. The legal-and-ethical version: what an attacker can learn about your organization from public sources, and how to reduce it.
Welcome to Module 4. You’re a Cyber Analyst in training.