Integrator Sort

Use this skill when a defense leader, program office, unit, or defense-tech company wants to decide who should integrate one class of system (industry, industry with warfighters judging, a Warwright team with reach-back, or a Warwright team at the unit).

This skill is Labor as Code™ for the sorting method in You Are Not the Warfighter. Labor as Code is a skill that encapsulates a person's knowledge, judgment, taste, and skill in durable, inspectable, versioned form. The know-how is treating "who integrates this?" as a question with an answer, decided from facts someone can write down, instead of a choice the first contract made by default.

The method spans four chapters. "Sort Every System by Who Should Integrate It" teaches the Integrator Sort, its four lanes, and its placement rules, worked on the Cairn Division's counter-drone kit as it grew from one kit to seven. "Warfighters Should Not Integrate What Industry Does Best" says where the Sort points to industry. "Pass Every Legal, Safety, and Cyber Gate Without Rebuilding the Old Chain" teaches the Gate Map, worked on the Larkspur Brigade's armed drone variant. "Start With One Builder, and Add Builders Only When One Can’t Do the Job" teaches Multi-Player Mode. Read them for the argument and the worked cases. This skill applies the method to the user's own system. It never borrows the cases' facts, costs, or dates as the user's.

The lane is a recommendation; government people decide it. The skill states which lane the evidence points to and why, because a Sort with no conclusion is an unfinished run. The lane itself is set by the government people who hold integrating authority, a commander or a portfolio acquisition executive (PAE), in writing. FAR 7.503 bars contractors from "determining what supplies or services are to be acquired by the Government." A contractor, a vendor, or this skill can supply facts for every answer. None of them owns the answer.

The skill never decides a gate. Whether a legal review, safety certification, airworthiness, weapons review, or cyber authorization applies is a legal and safety determination. The book gives the Gate Map no AI box for that reason: a list that misses a gate sends a team past a review it owed, and a list that invents one teaches the team to distrust the whole map. So the skill drafts the map from the book's full list, drops nothing on its own judgment, and marks every row "to be confirmed by the owning authority."

Before running

1. Apply the book's security rule first. Use only unclassified, publicly releasable material, in an AI tool the user's organization has approved for it. No capability, vulnerability, or operational detail: no frequencies, ranges tied to named systems, countermeasure specifics, tactics, unit locations, or schedules. The Sort runs on what the system is, what it's built from, how many there are, and what it connects to. If the user pastes anything marked CUI, FOUO, or classified, or anything that reads like operational detail, stop, say so, and ask for a releasable description. Nothing controlled goes into the AI Workforce™ running this skill: its files, connections, or memory.
2. Get one class, not one serial number. A class is a kind of thing fielded in some quantity. Sort a family of systems one class at a time.
3. Get a plain description. What it does; its parts, with makers and origin where known; how many exist and are planned; what network it connects to, if any; whether it carries or could carry a weapon; its weight class if it flies; how long it should serve; any test reports. Record each fact's source.
4. Ask who will sign, and who wrote the description. Name the government person who will own the Sort sheet. If a contractor wrote any of the description, record it; their facts stay, and the government signer restates the conclusions.
5. Ask what prompted the Sort: copies, a network connection, arming, a weight-class change, a funding request. The trigger usually tells you which answer just moved.

Step 1: Answer the seven questions with evidence

Answer each from the user's material only, and quote or cite the fact behind it. Where the material is silent, write "unknown" and name who could answer. An answer without evidence is a preference with a number on it.

1. Consequence. If it fails, who is hurt, and did they choose the risk: the crew or unit using it, other friendly forces, or civilians? Is the harm containable by test?
2. Gates. Which certifications must it pass, and can a unit-level team carry them, even with its partners? (Weapons legal review, airworthiness above Group 2, nuclear surety, energetics and explosives safety, an ATO on DoD networks, spectrum, component-origin rules.) Mark each "to confirm"; Step 3 lays out the full map.
3. Scale. How many are needed, and does producing them need rate production and a qualified supply chain?
4. Life. How long will it serve, and who sustains it after the builders rotate?
5. Tempo. How fast does the threat change what it must do: weeks, months, or years? A tempo with no source is "unknown."
6. Openness. Can it be built from components with documented interfaces the government has rights to use? Note origin here too.
7. Maturity. Are its component technologies mature, or is invention still required?

List every "unknown" in one place. It's the to-do list for the people who will sign the sheet, and often the most useful output of the run.

Step 2: Apply the placement rules and recommend a lane

The Sort doesn't average. Apply the rules in order.

1. Check the three trumps. Any one puts the class in lane I or II, whatever the other answers say: a consequence answer of "people who didn't choose the risk, and the harm isn't containable by test"; a gate a unit-level team can't carry, even with its partners; or a need for rate production. Name the trump and its evidence, and stop arguing speed or openness. Choose lane I for exquisite platforms, nuclear systems, manned aircraft and ships, munitions, and interceptors. Choose lane II for major subsystems on those platforms (radars, engines, weapons), where Warfighter Evidence Briefs feed the requirements, test at D0 to D1 decides acceptance, and the government still buys the data it needs.
2. If no trump fires, weigh the pulls. Fast tempo and open, documented components push toward lanes III and IV. Long life and larger scale push toward lane III, because a system serving for years across many units needs an organization that outlasts its builders. Lane IV fits classes that are short-lived, local, and cheap enough to throw away: attritable and consumable systems, printed parts and tools, sensor kits, plug-ins, training aids. Lane III fits mission software on existing platforms, planning and command-and-control applications, mission kits, tactical data integration, and systems fielded across many units, with reach-back to a software factory, depot, government lab, or PAE.
3. Name what would move it. More copies, a network connection, arming, a weight-class change. A lane IV prototype often becomes lane III the moment another unit copies it, because copying is scale.

Money isn't a Sort question, but it can force a rerun. If the user supplies an assembled cost, note whether it crosses the dated thresholds the book cites ($15,000 micro-purchase since October 1, 2025; the $350,000 FY2026 expense/investment line for an assembled system), marked "for the resource manager to confirm."

Step 3: Draft the Gate Map

At the top, write the system and configuration, the consequence answer from Step 1, and the latest change. Then one row per gate, every gate the book lists:

1. "Applies?" reads "to be confirmed by the owning authority" in every row. The skill may add why the description suggests a gate could apply, or that nothing in it touches the gate yet. It never writes "does not apply," and it never deletes a row.
2. Owner. Use the role the book names unless the user supplies a name: an authorized reviewing attorney, advised by the judge advocate; the weapons board; senior Department review; the service's airworthiness authority; the certifying authority in the unit's safety plan; the authorizing official (for a closed-loop drone, the O-6 under the July 10, 2025 memo); the spectrum manager; the explosives safety office; the judge advocate and contracting office; the resource manager; the contracting officer.
3. Clock. Only clocks the book states, dated, and "no clock" otherwise: the weapons board's 30 days and battery certification's one week under the July 10, 2025 memo; Group 1 and 2 drones exempt from airworthiness "with few exceptions," Group 3 back inside (each service's own implementation unconfirmed); the FY2026 NDAA's escalation path for ATO requests pending 180 days. Mark the autonomy row "being rewritten under NSPM-11; check the current version."
4. Evidence it needs. What the material holds, and "not in the material" for the rest.
5. Re-entry. The change that would send the build back through the gate: a network connection, a new battery lot or supplier, a custom part that changes origin status, a weight-class change, a new release mechanism, an autonomous feature. Size each gate to the consequence answer, and check consequences past the threat that started the project.
6. Flag every gate with no clock. It has to start first.

End the map by saying it goes to the judge advocate, the safety and test lead, the authorizing official, the resource manager, and the contracting officer together.

Step 4: Recommend a Multi-Player Mode for lanes III and IV

Multi-Player Mode names the scale a Warwright effort runs at, by how many people and AI Workforces are in play. An AI Workforce is a SpecKit, a harness, agents, skills, and MCPs, assembled toward one outcome and a defined set of deliverables.

• Single-player mode: one Warwright, one AI Workforce. Right for a bounded piece (a plug-in, a staff section's app, a configuration and its build record, a test card set, a parts log), never a whole class. It ships only inside a team or with a named reviewer, and its build record passes the Sustainment Tail Test's question 2 first.
• Team mode: a Warwright team of roughly four to twelve, most of them warfighters who fight with the system, sharing one specification, one harness, shared skills, and one build record. The default for a lane III or IV class owned end to end. Its failure: the shared AI Workforce quietly becomes one member's.
• Multi-team mode: several teams, each with its own AI Workforce, federated through shared specifications, documented interfaces, a shared registry, and a coordinating layer that reconciles and never decides. Right when the class belongs to a family fielded across a division or a fleet on one interface set. Its risk: the coordinating layer turns back into the old chain. Share interfaces and evidence, not one approval queue.

State the mode, the reason, and what it can't own. Every mode keeps its evidence at D0 or D1 on Distance to the Fight and passes the same gates. For lanes I and II, no Warwright mode integrates; warfighters judge at D0 to D1.

Step 5: Conclude with a lane recommendation

Close with a short statement, not an evidence pack:

1. The recommended lane, in one line.
2. The reasoning, in three to five sentences tied to specific answers: which trump fired or didn't, and which pulls decided it.
3. What would move it, and when to re-run: whenever quantity, network connection, armament, or weight class changes, and before anyone asks for money to build copies.
4. The next action. For lane I or II, tell the unit in writing so nobody spends a year building in the wrong lane. For lane III or IV, name the team that would integrate it and the commander or PAE who would grant integrating authority in writing; the Warwright Team Charter in "A Warwright Team Needs Authority, Money, and Time Before It Needs Code" says what that grant contains.
5. Who signs: the government person from "Before running this."

Where answers conflict and no trump settles them (tempo pulling toward lane IV against life and scale pulling toward lane III, or a "to confirm" gate that would decide between lanes II and III), present the live options side by side: each lane, its reasoning, the answer that would make it right, and who can supply that answer. Where too many answers are "unknown" for any lane, say so, list the unknowns, and name who fills each. A lane built on answers with no evidence beside them should be re-run before anyone buys anything.

Checks before calling the work done

• Every answer cites the user's material or reads "unknown." Delete any fact supplied from general knowledge of a system type.
• The trumps were checked first, and none was argued away by tempo or openness.
• All eleven Gate Map rows are present, each marked "to be confirmed by the owning authority."
• No capability, operational, or controlled detail appears.
• The conclusion names a lane or labeled options, with reasoning, what would move it, and who signs.

What this skill does not do

It doesn't grant integrating authority, contracting authority, or money. It doesn't decide whether a gate applies, whether a change is significant, or whether a weapon is lawful; the reviewing attorney, the authorizing official, and the safety and airworthiness authorities do. It doesn't let a contractor's paragraph, or its own, stand in for a government decision about what to acquire. And it doesn't argue a class out of industry's lanes because a unit could build it faster.

Used in Acquired Alpha

A buyer deciding who should integrate a class of systems is making a purchase of a capability. In Acquired Alpha the sort's answer is one input to the Three-Way Sort, and each claim an integrator makes about its team goes through the book’s claim test (the Three Tests). The skill's own conclusion does not change.