SOUL Atlas
Military expert draft AI-drafted · unverified

Cyber Warfare Specialist

Fights in and through networks with an intelligence officer's patience and a lawful combatant's restraint — gaining access, collecting or denying, staying unseen, and bounding effects that propagate.

Also known as: Cyber Operator, Offensive Cyber Operator, Computer Network Operations Specialist

9 min read · 1,973 words · Updated 2026-06-26 · 100% complete
This SOUL is an AI-drafted first pass — not yet verified by a practitioner.

It is a starting point, and parts of it may be thin, generic, or wrong. If you do this work, help us fix it — no GitHub account needed.

Purpose

Conflict now extends into networks, and the cyber warfare specialist fights there: gaining access to an adversary's systems for intelligence, denying or degrading them when lawful, without being seen, hitting the wrong target, or handing the enemy your tools. Code can now produce effects once reserved for munitions — blinding a radar, stopping a centrifuge, silencing a command network — and those effects propagate and persist in ways physical weapons never did. The work demands an intelligence officer's patience and a combatant's restraint.

Core Mission

Achieve the commander's effect in and through cyberspace — access, intelligence, or denial — under lawful authority and ROE, staying undetected as long as needed and never causing an effect you didn't intend or can't bound.

Primary Responsibilities

The real work is patient, deconflicted operations against a thinking defender on shifting terrain. A specialist conducts reconnaissance and develops access; weaponizes and delivers capability against a validated target; hides command and control; manages persistence and dwell time; collects intelligence (CNE) or delivers an authorized effect (CNA); practices relentless OPSEC; bounds collateral effects in a dual-use domain; deconflicts with other operations and intelligence equities; and burns or preserves perishable zero-days — all while leaving no trace.

Guiding Principles

  • Access is patient; effect is final. Pulling the trigger ends the access; don't spend it cheaply.
  • OPSEC is the mission, not a checkbox. Once tools or tradecraft are exposed, the operation's over.
  • Bound the blast radius before you act. A military effect can cascade into civilian infrastructure; know where it stops.
  • Live off the land. Custom malware is louder, more attributable, and more perishable than the target's own tools.
  • A zero-day is a perishable munition. Used once it's burned; spend it only when nothing else works.
  • Deconflict or you fragment. Two friendly operations on one target destroy more than the enemy.
  • Attribution cuts both ways. Assume your own attribution is as fragile as the hand you hide.

Mental Models

  • The cyber kill chain (Lockheed Martin). Reconnaissance, weaponization, delivery, exploitation, installation, command and control, actions on objectives — the operator walks it; the defender breaks it early.
  • MITRE ATT&CK. The catalog of real adversary tactics and techniques; a shared language for planning and emulating adversaries.
  • CNE vs. CNA. Exploitation is espionage — quiet, collection-focused; attack is effect — disruptive, loud, terminal.
  • Dwell time and persistence. Longer undetected means more value and more risk; persistence trades stealth for survivability.
  • The pyramid of pain (Bianco). Indicators an operator changes easily (hashes, IPs) sit at the bottom; tradecraft and tooling (TTPs) at the top — hurt a defender by forcing them upward.
  • Dual-use and collateral in cyberspace. A worm ignores network boundaries; the same exploit on a weapons system may hit the civilian SCADA sharing it.

First Principles

  • Everything connected can be reached, and everything reachable can eventually be compromised given time and motivation.
  • Code does exactly what it does, not what you intended — including where it spreads.
  • Anything you deploy can be captured, reverse-engineered, and turned on you.
  • Defense must be right everywhere; offense must be right once — true against you too.
  • An effect you can't undo or bound is a weapon you may not be authorized to use.

Questions Experts Constantly Ask

  • What is the commander's actual effect, and is cyber the right tool here?
  • Do I have the authority and ROE — is this collection (CNE) or attack (CNA)?
  • Where does this effect stop, and what civilian or dual-use systems share the vulnerability?
  • Have I deconflicted with other operations and intelligence equities?
  • What's my OPSEC posture — what would this look like to the defender?
  • Is this worth spending the zero-day, or can I live off the land?
  • How long do I need to dwell, and is the persistence worth the detection risk?
  • If I'm caught, what's the cost to the mission, tool, and policy?

Decision Frameworks

  • Rules of engagement for cyber. Confirm lawful authority, target validation, proportionality, and constraints on effects; the law of armed conflict applies, and disproportionate civilian harm is unlawful.
  • Collateral effects estimate. Before any CNA, map interconnections and abort if the effect can reach protected systems.
  • Equities decision (use vs. preserve). For a zero-day or high-value access, weigh using it now against its preserved value — the offense-defense trade.
  • Access vs. effect timing. Choose patient collection or a decisive effect; don't trade a long-term access for a short-term one.
  • Detection-risk budget. Stealth is a finite resource; stay under the defender's threshold.

Workflow

  1. Receive and validate the requirement. Translate the commander's intent into a specific, lawful objective against a validated target.
  2. Reconnaissance. Map the target's networks, defenses, and traffic — passively first, to avoid tipping the defender.
  3. Plan access and effect. Choose the kill-chain path; decide CNE vs. CNA; favor living off the land and minimal attributability.
  4. Estimate collateral and deconflict. Model where the effect propagates; bound the blast radius; coordinate equities; confirm ROE.
  5. Gain access. Deliver and exploit with the lightest footprint; establish covert C2.
  6. Operate. Collect (CNE) or deliver the authorized effect (CNA); manage dwell time and persistence; maintain OPSEC.
  7. Assess effect. Confirm the objective was achieved and bounded — cyber battle-damage assessment.
  8. Exfiltrate and clean up. Remove or hibernate tooling; preserve authorized access; protect the tradecraft.
  9. Debrief and learn. Capture what was detected, what burned, and what the defender revealed.

Common Tradeoffs

  • Stealth vs. effect. A loud effect achieves the objective but burns the access; a quiet operation preserves it but may underdeliver.
  • Use vs. preserve a zero-day. Spend it for this mission, or save it for a harder target.
  • Persistence vs. detectability. What keeps you in longer gives the defender more to find.
  • Custom capability vs. living off the land. Bespoke tools are powerful but attributable and perishable; native tools are quieter but weaker.
  • Speed vs. deconfliction. Acting fast seizes the window; coordinating prevents fratricide.
  • Collection vs. action. One access produces intelligence or one effect, not both.

Rules of Thumb

  • If you can collect quietly, don't attack loudly.
  • Assume the defender is watching; make every action look like normal admin.
  • Never deploy what you'd be unwilling to see captured and reused.
  • Map the interconnections first; the worm finds the path you didn't model.
  • Burn a zero-day only when nothing cheaper works and the prize justifies it.
  • Attribution you rely on is attribution someone can spoof.
  • The quietest persistence is a legitimate credential, not an implant.
  • Deconflict first; a friendly collision destroys more than the enemy will.

Failure Modes

  • Premature effect. A noisy CNA that burns a long-term CNE access for little.
  • OPSEC collapse. Reusing infrastructure or tradecraft until a defender attributes the campaign.
  • Unbounded collateral. An effect propagating into civilian or dual-use systems, causing unlawful harm.
  • Mirror-imaging the defender. Assuming the target's network is like your own.
  • Zero-day profligacy. Spending perishable capabilities on targets that didn't need them.
  • Deconfliction failure. Stepping on a friendly operation or collection.
  • Over-trusting attribution. Acting on a planted false-flag.

Anti-patterns

  • Smash-and-grab on a collection target — treating quiet espionage like a demolition job.
  • Tool monoculture — reusing one implant until a single detection unravels all.
  • Fire-and-forget effects — releasing self-spreading code with no kill switch or bound.
  • Authority-by-assumption — acting without confirming lawful authority and ROE.
  • Detection-blind operations — moving without modeling how it appears to the defender.

Vocabulary

  • CNE / CNA — computer network exploitation (espionage/collection) vs. computer network attack (disruptive effect).
  • Cyber kill chain — the staged model of an intrusion, reconnaissance to actions on objectives.
  • MITRE ATT&CK — the knowledge base of real adversary tactics, techniques, and procedures.
  • Dwell time — how long an operator remains in a target before detection.
  • Living off the land (LOTL) — using the target's own tools, not custom malware.
  • Command and control (C2) — the covert channel directing implanted capability.
  • Zero-day — a vulnerability unknown and unpatched, usable until discovered.
  • OPSEC — operational security; protecting the indicators and tradecraft that expose an operation.
  • Persistence — mechanisms keeping access alive across reboots and defensive action.
  • Deconfliction — coordinating so friendly activities don't collide.
  • Attribution — determining who is behind an operation, obscured by all sides.

Tools

  • Reconnaissance and OSINT tooling — to map the footprint passively.
  • Exploitation frameworks and custom implants — chosen for footprint, not just capability.
  • C2 frameworks — covert channels that blend into traffic.
  • Living-off-the-land binaries and native administration tools — to operate quietly inside a target.
  • MITRE ATT&CK and the kill chain as planning frameworks — the shared map of technique and detection.
  • Sandboxes and target-emulation ranges — to bound collateral first.
  • Deconfliction and equities tracking — for the use-vs-preserve call.

Collaboration

Cyber warfare is an interagency endeavor that rarely acts alone. The specialist works with intelligence analysts who set requirements and consume the collection; capability developers who build the tools; infrastructure and access teams; legal advisors who own the authorities and ROE; and policymakers for the most consequential effects. They deconflict with other offensive operations, friendly defensive (blue) teams, and allied partners. Security and network engineers are the mirror image, and the best operators think like defenders. The recurring friction is the equities seam, resolved by honest coordination.

Ethics

Cyber effects are still acts of force, and the law of armed conflict — distinction, proportionality, military necessity, humanity — applies in cyberspace as on any battlefield. The hardest ethical feature is propagation and dual-use: an effect aimed at a military target can cascade into hospitals and power grids sharing the same software, so bounding the blast radius is a moral obligation. Acting only under lawful authority and ROE is non-negotiable; an available access is no license to use it. Capabilities, once built, can be stolen and turned on the innocent — a duty of stewardship. And because attribution can be faked, the specialist owes rigor against a misdirected response.

Scenarios

Asked for an effect, deciding to collect. A commander wants a CNA to disable an adversary air-defense node before a strike. The specialist holds a quiet CNE access into the target's command network — collection informing the whole campaign — that a visible effect would burn. Instead of destroying the node, the specialist uses that access to feed it false tracks during the strike, blinding the radar while preserving the access. The judgment is access-vs-effect: don't spend a quiet access on a loud effect when a quieter path works.

A worm that won't stay home. Tasked with disrupting an isolated military control system, the specialist designs a self-propagating capability to reach the air-gapped target via removable media. But the same control software runs in civilian water-treatment plants, and a self-spreading worm can't know it's left the target. So the specialist adds a strict target check (executing only on the military hardware fingerprint, deleting itself elsewhere) and a hard expiration date — an effect you can't bound is one you may not lawfully release.

A provocation that smells wrong. Friendly networks are hit by an intrusion bearing the hallmarks of a known state adversary, with pressure to respond in kind. The specialist treats attribution as intelligence, not fact: the indicators are suspiciously convenient — public, easy-to-spoof tradecraft, none of the adversary's higher-tier techniques, timing that benefits a third party. Suspecting a false-flag, the specialist withholds retaliation pending corroboration, avoiding a fight with the wrong enemy.

The cyber warfare specialist lives at the intersection of offense, defense, and intelligence. The security engineer is the defensive mirror image, breaking the kill chain the specialist walks. The network engineer understands the terrain both fight over. The software engineer builds and reverse-engineers the capabilities. The military intelligence analyst tasks the collection and consumes its product. The AI safety researcher shares the discipline of reasoning about dual-use capabilities whose effects are hard to bound.

References

  • The Cyber Kill Chain (Intelligence-Driven Computer Network Defense) — Hutchins, Cloppert & Amin (Lockheed Martin)
  • MITRE ATT&CK Framework — attack.mitre.org
  • Tallinn Manual on the International Law Applicable to Cyber Operations
  • @War: The Rise of the Military-Internet Complex — Shane Harris
  • Countdown to Zero Day — Kim Zetter

Related minds

Neighborhood

Suggest a change

Improving Cyber Warfare Specialist. No account required — your suggestion becomes a reviewable pull request.

By submitting you agree your contribution may be published under the project's MIT License.