# CORNERSTONE
## A national framework for correctional campuses

**Born Between 2 Generals, LLC**
Framework document, version 1.0 — July 2026

---

## The one-page version

Roughly 95% of people in American prisons come home. The question was never whether they return to our communities. It is what version of them arrives.

CORNERSTONE turns the time between sentencing and release into the thing it should always have been: a period of structured training that produces a person with a credential, a job offer, treated clinical need, and a verifiable record — instead of a person with a release date and nothing else.

The evidence base is not contested. RAND's meta-analysis of correctional education found participants had **43% lower odds of returning to prison**, a 13 percentage-point reduction in recidivism risk, and a return of **$4 to $5 saved for every $1 invested** over three years. Vocational participants were 28% more likely to be employed after release.

**The framework has five parts:**

0. **The floor** — the conditions of confinement that everything else stands on. A campus cannot be built inside a building that is making people sick. This comes first, not last.
1. **The campus model** — a five-tier program architecture that a facility can actually staff and fund.
2. **The funding stack** — existing federal streams, matched tier by tier. No new appropriation required to start.
3. **The national strategy** — four parallel lanes, because corrections is roughly 85% state and local and cannot be mandated from Washington.
4. **The separation framework** — three independent axes replacing the single offense-label exclusion list that sinks most reform proposals.

**And five working applications** — FOOTING, KEYSTONE, BEARING, PARCHMENT, and CHALKLINE — that make the framework operable rather than theoretical.

---

# Part 0 — The floor

## Why this section comes first

Everything in this document after this page assumes a building where a person can breathe.

That assumption does not currently hold. People are dying inside American correctional facilities from environmental conditions — mold, ventilation failure, water, untreated heat — and from medication that is prescribed and then simply not delivered. These are not program-quality problems. They are habitability problems, and no amount of curriculum design fixes a room that is poisoning the person sitting in it.

There is also a strategic reason this belongs at the front rather than in an appendix. **A rehabilitation framework that opens with college programs while people are getting sick in the same building will be read as unserious, and it will deserve to be.** The credibility of everything downstream depends on the floor being addressed first and explicitly.

## The standard

A facility qualifies as a correctional campus only if it meets a habitability floor, verified by an entity that does not report to the corrections department:

1. **Air.** Independent air-quality and mold inspection on a published schedule, with results public and remediation deadlines enforceable.
2. **Water.** Potable water tested to the same standard as any other public building.
3. **Temperature.** Enforceable maximum and minimum interior temperatures, with medical vulnerability accounted for.
4. **Medication continuity.** Prescribed medication delivered without interruption, measured as a rate, published, and carrying contract penalties when missed.
5. **Timely medical access.** Defined maximum wait from request to clinical assessment, published and audited.
6. **Freedom from sexual and physical abuse.** Full PREA compliance verified by outside audit, with a reporting channel that does not run through the facility.
7. **Sanitation.** Functioning plumbing, pest control, and clean linen and clothing on a schedule.
8. **A working grievance system.** Numbered, tracked, answered within a published window, and audited for retaliation.

## The two structural failures underneath all of it

**Health department jurisdiction is patchy.** In a number of states, correctional facilities are wholly or partially exempt from the routine public health inspection that governs restaurants, schools, and nursing homes. A building the state would close if it were a daycare stays open because of who lives in it. **Ending that exemption is one of the highest-leverage and most winnable legislative fixes in this entire document** — it costs almost nothing, it is hard to argue against out loud, and it creates an independent inspection record that changes every subsequent fight.

**Independent oversight is missing.** Roughly half of states have no corrections ombudsman or independent oversight body with real access. Where the only entity investigating a facility is the department that runs it, conditions problems are managed as reputational problems.

## The Eighth Amendment position

Conditions of confinement are governed by the deliberate indifference standard: *Estelle v. Gamble* for medical care, *Farmer v. Brennan* for failure to protect.

For environmental exposure specifically, **the case is *Helling v. McKinney* (1993)**, which held that exposure to conditions posing an unreasonable risk of *future* harm can violate the Eighth Amendment even before anyone has become ill. For a mold claim this is the controlling framework, and it means the argument does not have to wait for another death to be made.

The practical obstacle is not the standard. It is the **Prison Litigation Reform Act's exhaustion requirement**, which bars most federal claims unless every level of the facility's internal grievance process was completed on time. This is where meritorious cases die — not on the merits, on a missed appeal deadline. Any conditions strategy that does not put the grievance clock at the center of it will lose cases it should have won.

## What this adds to the legislative package

Three components, all of which belong in the same state bill as the campus authorization:

- **End the correctional exemption from public health inspection.** Same standard, same inspectors, same publication requirement as any other institutional building.
- **Establish independent corrections oversight** with statutory access, subpoena power, and public reporting.
- **Medication continuity as a contract standard**, measured and penalty-bearing, in every correctional healthcare contract.

## The parallel channel that nobody uses

If air quality is sickening incarcerated people, it is sickening the correctional officers breathing the same air.

**An OSHA complaint filed by a staff member creates an independent federal inspection record about that building** — one that the corrections department does not control, cannot characterize away, and cannot resolve internally. One officer or one union willing to file changes the entire evidentiary picture of a conditions case, and it is almost never used.

---

# Part I — The model

## Two design corrections that come before everything else

### 1. Jails and prisons are different problems

This is the correction that separates a serious proposal from a dismissible one.

**Jails** are roughly 3,000 county-run facilities. Most people in them are pretrial and not convicted of anything. Median stays are measured in weeks. There is no version of a degree program that works here — and enrolling someone who will be released mid-course produces a failed transcript, which is worse than no transcript.

**Prisons** are roughly 1,100 state and federal facilities where people stay for years. This is where the campus belongs.

So the model splits:

| | Jails | Prisons |
|---|---|---|
| **Runs** | Tier 0 and Tier 1 only | Full Tier 0–4 campus |
| **Priority** | Stabilization, assessment, identity documents, benefits enrollment | Credentials, degrees, apprenticeship hours, clinical treatment |
| **Success metric** | Warm handoff to a community provider running the same plan | Credential completion and employment at 12 months |

Anything that calls the whole thing "jail reform" will be corrected by every corrections professional in the room. The term is **correctional campus**.

### 2. Sort by risk and need, not by offense

The strongest evidence base in corrections is Risk-Need-Responsivity: you allocate programming by assessed *risk of reoffending* and *criminogenic need*, not by conviction category.

Sorting by offense label is how systems end up delivering the most programming to the lowest-risk people and none to the highest-risk ones — precisely inverting the intervention.

Offense category still matters enormously. It just belongs on a different axis. See Part IV.

---

## The campus: five tiers

Order here is real, so the numbering carries information.

### Tier 0 — Intake and assessment (60–90 days)

Literacy and numeracy testing, learning-disability screening, trauma and substance-use screening, validated risk and need assessment, aptitude battery, work history.

Output: **one individualized plan that follows the person through every transfer.** Not a file that gets rebuilt at each facility. One plan, one record.

Assessment is the only thing that must be right before anything else can be built. It is also the cheapest tier and the one most often skipped.

### Tier 1 — Foundation

Literacy, high school equivalency, English language instruction, and the cognitive-behavioral core (Thinking for a Change, Moral Reconation Therapy or equivalent).

This is a non-negotiable floor. A substantial share of the incarcerated population reads below a sixth-grade level. Nothing in Tier 2 works until this is addressed, and every dollar spent above an unaddressed literacy deficit is wasted.

Tier 1 is also the single largest driver of institutional safety. Facilities where people are in classrooms are facilities where fewer people are assaulted — staff included.

### Tier 2A — Academic track

Accredited college coursework through an approved Prison Education Program, delivered by a public or nonprofit institution. Stackable credentials: certificate → associate → bachelor.

Funded by Pell. The 1994 ban was lifted by the FAFSA Simplification Act; incarcerated students became eligible again as of July 1, 2023, through programs approved under 34 CFR 668.236. For-profit institutions are statutorily barred from operating an eligible PEP.

### Tier 2B — Trades track

**Equal standing with the academic track. Not a consolation tier.**

- DOL Registered Apprenticeship with on-the-job hours that count toward journeyman status on the outside
- Portable, industry-recognized certifications: NCCER, AWS welding, EPA 608, ServSafe, CompTIA A+/Network+/Security+, CDL where state law permits
- Journeyman mentors drawn from long-sentence populations where appropriate

**The failure point most proposals miss:** occupational licensing. You can train a flawless welder, barber, electrician, or nurse aide and the state licensing board will refuse them at the door on a blanket conviction bar. **Licensing reform must be in the same bill.** A credential the licensing board will not honor is not a credential; it is a certificate of frustration.

### Tier 3 — Clinical, targeted by assessed need

Universal "extensive therapy" is not achievable. There are not enough licensed clinicians in the country, and promising it makes the whole proposal sound naive.

What is achievable, in priority order:

1. **Medication-assisted treatment** for opioid use disorder. This has the largest immediate effect of anything in the model — post-release overdose is a leading cause of death in the first two weeks after release, and MAT cuts it sharply.
2. **Trauma-focused therapy** allocated by assessment.
3. **Structured violence-reduction programming** for the high-risk violent population, which is where the recidivism numbers actually live.
4. **Offense-specific treatment** where indicated — added on top of the general tiers, never substituted for them.

### Tier 4 — Transition (final 18 months)

Work release. Employer pre-hire commitments signed before the release date, not sought after it. Housing secured. Identity documents in hand. Bank account opened. Benefits enrolled before the gate.

**Most reentry failure is logistics, not motivation.** A person released with a degree, no ID, no address, and $40 fails for reasons that have nothing to do with rehabilitation.

For life-without-parole populations, Tier 4 is structurally inapplicable — substitute the peer-mentor and instructor track, which is where long-sentence populations do the most good and where the campus gets its most stable staff.

---

# Part II — The money

Nothing here requires a new appropriation to begin. These are existing streams, matched to the tier each can legally pay for.

| Stream | Covers | Notes |
|---|---|---|
| **Pell — Prison Education Program** | Tier 2A tuition and fees | Requires approved PEP; nonprofit or public institutions only |
| **Workforce Pell** | Tier 2B short-term credentials | Created by the 2025 reconciliation act, extending Pell to short-term job training. **Newest lever in the stack and the least used anywhere. This is the opening.** |
| **WIOA Title II (AEFLA)** | Tier 1 adult basic education and HSE | Corrections is a permitted set-aside |
| **Perkins V** | Tier 2B equipment, curriculum, instructors | Standard CTE funding |
| **DOL Registered Apprenticeship** | Tier 2B sponsor build-out, mentor stipends | Also provides the hour-tracking infrastructure |
| **Medicaid 1115 reentry waiver** | Tier 3 clinical and MAT, final 90 days pre-release | Requires approved state waiver — the single highest-leverage state action |
| **Second Chance Act** | Tier 4 transition, housing navigation | Long-standing, competitive |
| **Justice Reinvestment** | State match, capital | Where states capture and redirect avoided incarceration cost |

**Workforce Pell deserves separate emphasis.** It is new enough that implementation guidance was still developing at last check, and no state has been identified with an operating corrections program using it. The first state to move here gets a federally funded trades track while everyone else is still reading the guidance. That is a first-mover advantage with a closing window, and it is the single most actionable item in this document.

---

# Part III — The national strategy

Corrections is roughly 85% state and local. "Nationwide" cannot mean a federal mandate. It means **federal demonstration, state adoption, and a standard everyone converges on.**

Four lanes, run in parallel. None of them waits for the others.

## Lane 1 — Federal demonstration

The Bureau of Prisons is the one system Washington controls directly. It is the proof site.

**The ask:** expand First Step Act earned-time credits so they attach to **credential completion**, not merely program hours. Today a person can log hours and earn credits without finishing anything. Tying credit to completion changes behavior on both sides of the desk — it gives the incarcerated person a reason to finish and the facility a reason to run programs that can be finished.

This is a small, scoreable, low-cost legislative ask with an existing statutory vehicle. Start here because it is winnable.

## Lane 2 — Anchor states, deliberately bipartisan

Pick five states with contrasting politics and run them simultaneously:

- **Texas** — the keystone. Right on Crime made correctional reform a conservative fiscal-responsibility issue here, and Windham School District is the longest-running correctional education system in the country.
- **Utah, Georgia** — conservative proof points at smaller scale.
- **California, Michigan** — the other direction. California is already spending real money: a $239 million Nordic-model education center at San Quentin opened in early 2026 under the California Model initiative, which drew on Norway's approach through the Amend program at UCSF.

Proof from both directions is what kills the "soft on crime" attack. One-party proof is not proof; it is a talking point.

## Lane 3 — Own the standard

**This is where the most leverage sits and the least competition.**

Build the accreditation framework itself: what a correctional campus must offer, what outcomes it must publish, how credentials must transfer, what assessment is required at intake. Publish the standard. Certify against it.

The American Correctional Association is the incumbent accreditor and is widely criticized as a paper exercise that certifies process rather than outcome. There is an opening for a credible outcome-based standard, and whoever writes it sets the national agenda without needing a single vote in a single legislature.

This is also the lane a nonprofit can occupy immediately, with no statutory change required.

## Lane 4 — Employer coalition

The demand side, and the reason the politics hold.

Construction, manufacturing, logistics, skilled trades unions, and utilities all have documented labor shortages. Get signed hiring-pipeline commitments — not statements of support, commitments with headcount attached.

An employer coalition converts the proposal from a request for mercy into a workforce solution. It also gives every legislator a constituent business asking them to vote yes.

## Sequencing

| Window | Action |
|---|---|
| **Months 0–6** | Publish the standard (Lane 3). File the FSA credential amendment (Lane 1). Open employer coalition recruitment (Lane 4). Launch CHALKLINE research pass 2. |
| **Months 6–18** | Two anchor states introduce legislation. First facility blueprints under KEYSTONE. Licensing-barrier bills filed in the same sessions. |
| **Months 18–36** | First certified campuses operating. PARCHMENT credential ledger live at pilot sites. First outcome data published. |
| **Years 3–5** | Standard adopted as a procurement condition in multiple states. Federal demonstration results published. Scale. |

## Framing — the part that decides whether it passes

**Never lead with redemption.**

Lead with the arithmetic: 95% come home. You are choosing between a person with a welding certificate and a job offer, and a person with neither. That is the entire choice, and it is a public safety choice before it is anything else.

| Say this | Not this |
|---|---|
| Public safety | Second chances |
| Workforce and labor shortage | Rehabilitation |
| Taxpayer return on investment | Compassion |
| Victims are safer when reoffending falls | Everyone deserves a chance |
| Accountability includes producing a person who can hold a job | Prison is inhumane |

The second column is not wrong. It simply loses, and it loses the votes of people whose support you need. The Texas experience proved that conservatives will pass this at scale when it is framed as fiscal responsibility and victim safety.

---

# Part IV — The separation framework

You asked for a way to separate out the people who cannot be rehabilitated in the near future. That instinct is sound and the framework needs it. But the two categories most people name first are the two where the data will surprise you — and you do not want to learn that from a hostile witness in a committee hearing.

## Two facts to build around

### Homicide

**People convicted of murder have among the lowest rearrest rates of any offense category.** This is one of the most consistently replicated findings in corrections research. Most homicides are situational rather than predatory, sentences are long, and people are released at older ages when criminal behavior has fallen off sharply.

An exclusion list built on the word "murderer" will be dismantled in a single paragraph by the first criminologist who reads it — and it will take the credibility of everything around it down too.

**What actually predicts:** age at release, pattern of prior violence, institutional conduct, assessed risk score. Build on those.

### Sex offenses against children

Contact sex-offense recidivism is lower than public perception, and offense-specific treatment delivered on risk-need-responsivity principles shows real if modest effects. The evidence is mixed, not empty.

What *is* settled is that the risk-management obligation is lifelong.

**"Untreatable" and "requires permanent supervision" are different claims, and only the second is defensible.** Make the second claim. It gets you everything the first claim would have, and it survives cross-examination.

## The three axes

Replace the single exclusion list with three independent decisions. One case produces three different answers, and that is the point.

### Axis 1 — Placement: who does this person live near?

Standard custody classification, driven by offense severity, remaining sentence, institutional conduct, escape history, and age.

Separate housing for people with offenses against children is not controversial — it is already routine practice, and it protects the person as much as anyone else.

**Design requirement:** custody level must be able to fall over time with conduct. Sentence length is not behavior. A reclassification interval of 12 months on clean conduct is the standard to write in.

### Axis 2 — Program access: what may this person learn?

**Near-universal, by design — including for people who will never leave.**

This is the counterintuitive one, and it is the safety argument rather than the mercy argument:

- Idle prisons are violent prisons. Programming is the single largest driver of reduced institutional violence.
- Long-sentence populations are the stabilizing element in a facility and the mentor bench the whole campus runs on.
- Staff assault rates fall where programming is dense.

An exclusion on Axis 2 must be **structural** — the program cannot physically apply, as with Tier 4 transition on a life-without-parole sentence — never punitive.

Offense-specific treatment **adds** to the tiers. It never substitutes for education or trades.

### Axis 3 — Release-related benefits: does this person go home sooner?

**This is where your separation strategy belongs, and where the carve-outs already exist in law.**

Federal template: 18 U.S.C. § 3632(d)(4)(D) lists approximately 68 offense categories that disqualify someone from earning First Step Act time credits — generally violent offenses, sex and sexual exploitation offenses, human trafficking, terrorism, espionage, and leadership-role drug offenses. Note also that the BOP treats a single disqualifying count in an aggregated sentence as disqualifying the whole sentence.

Everything you want to exclude — earned time, parole consideration, work release, community custody — lives here. Put it here, and Axis 2 stays open, and the proposal survives contact with the evidence base.

### The fourth question — genuinely undischargeable risk

For the narrow population where risk genuinely does not discharge at the end of a sentence, the existing mechanism is **post-sentence civil commitment** for sexually violent predators. It exists in roughly twenty states and federally under the Adam Walsh Act (18 U.S.C. § 4248). It was upheld in *Kansas v. Hendricks* (1997) and *United States v. Comstock* (2010).

It is expensive, it rarely releases anyone, and it is constitutionally contested. Know that it is the existing tool and know its problems before proposing anything to replace it — because that will be the first question you are asked.

## The constitutional floor

**You cannot detain someone past their sentence for being "unrehabilitated."** *Foucha v. Louisiana* (1992) forecloses it.

Any preventive detention beyond a completed sentence must be:

- **Civil rather than penal** in character
- Tied to a **diagnosed condition plus present dangerousness**, not to the offense alone
- Subject to **periodic review** with a genuine release valve
- Supported by clear and convincing evidence at minimum

Design around this on day one. It is the first thing opposing counsel will reach for, and a framework that ignores it dies in court no matter how good the program design is.

---

# Part V — The legislative package

## Federal

1. **FSA Credential Completion Amendment** — attach earned-time credits to completed credentials rather than logged hours. Small, cheap, winnable, and it is the demonstration.
2. **Workforce Pell corrections guidance** — push ED for explicit guidance on short-term-credential Pell inside correctional settings. This may not need legislation at all; it may need a letter and a meeting.
3. **PEP approval streamlining** — the approval pathway runs through oversight entities, accreditors, and ED, with a best-interest determination at the end. Every month of delay is a cohort that does not enroll.

## State model bill — nine components

A state that passes all nine has a campus. A state that passes eight has a program that will underperform and be blamed for it.

**The floor — components 1 through 3. These are the easiest to pass and should lead the bill, not trail it.**

1. **End the correctional exemption from public health inspection** — correctional facilities inspected on the same standard, by the same inspectors, with the same publication requirement as any other institutional building.
2. **Independent corrections oversight** — an ombudsman or oversight body with statutory access, subpoena power, and public reporting, not reporting to the department it oversees.
3. **Medication continuity as a contract standard** — delivery measured as a published rate, with penalty provisions in every correctional healthcare contract.

**The campus — components 4 through 9.**

4. **Correctional campus authorization** — establishes the tier structure and the assessment requirement at intake.
5. **Occupational licensing reform** — prohibits blanket conviction-based denial; requires a direct-relationship test and written reasons. *Non-severable from component 4 — they pass together or the trades track is decorative.*
6. **Medicaid 1115 reentry waiver directive** — instructs the state Medicaid agency to seek or expand a reentry waiver. This is the clinical funding line.
7. **Earned-time for credential completion** — the state analogue of the federal ask.
8. **Portable credential record mandate** — requires that education and training records be issued in a transferable, verifiable format that survives facility transfer and release. This is the PARCHMENT standard.
9. **Outcome publication requirement** — the department must publish enrollment, completion, employment at 12 months, and recidivism at 36 months, disaggregated. Without this, nothing can be evaluated and the program is defenseless at its first budget hearing.

---

# Part VI — The CORNERSTONE app suite

Five applications, two shared core modules — one for the design system and data handling, one for accessibility. Every app carries read-aloud, adjustable text size, high contrast, and English/Spanish switching, because the people who most need FOOTING are often reading on a phone, in a second language, at the worst moment of their year.

Five applications. Owned by Born Between 2 Generals, LLC and licensed to qualifying nonprofits and public agencies.

### FOOTING — conditions reporting for families
The floor, made operable. A family member on the outside enters what is happening — mold, water, medication not arriving, sexual assault, retaliation — and gets back every office with authority over it, how fast each one moves, and which ones do not require naming the person inside.

Three things make it different from a complaint form:

- **The safety brief comes before the form.** Retaliation lands on the incarcerated person, not the reporter, and the app says so first.
- **The routing fan.** Families are told there is one door. There are nine, and several are invisible from the outside — the state medical licensing board, which follows a prison clinician personally regardless of who employs them; healthcare contract oversight, where missed medication passes are a measurable penalty-bearing breach; and the OSHA channel described above.
- **The grievance clock.** Tracks PLRA exhaustion from the outside, because the person inside may be too sick, too frightened, or too isolated to track it themselves. This single feature probably saves more claims than everything else in the app.

**Jurisdiction layer — all 51 built.** Every state plus the District of Columbia, eleven fields each, each marked confirmed, likely, or open: department of corrections, state health agency, protection and advocacy agency, medical board, nursing board, ACLU affiliate, independent oversight body, health-department jurisdiction status, grievance deadlines, healthcare contractor and contract oversight, and state PREA coordinator.

Pass 1 status: **561 cells — 322 carrying verified agency names, 239 open.** Independent corrections oversight bodies identified in 16 of 51 jurisdictions, which is itself one of the more damning findings in this document: in most of the country there is no entity investigating a facility except the department that runs it.

**One deliberate constraint: the directory carries agency names and no contact details.** No phone numbers, no web addresses, none generated. A wrong number given to a family during a medical emergency is worse than no number, so contact verification is a research-queue task with a named verifier and a date, not something inferred. The queue exports as CSV with a verification column per cell, which makes it delegatable to volunteers without handing them judgment calls.

**Boundary note:** FOOTING is family-facing and routes to outside jurisdiction. It does not replace or overlap The Watch or Witness in the Angela Stanton King suite, which are the insider-facing pair. Different reporter, different channel, different threat model.

### KEYSTONE — campus blueprint builder
Enter a facility's population, length of stay, and profile; get tier allocation, staffing at defined ratios, a cost and return model built on RAND's figures, a funding stack matched tier by tier, and a 36-month build sequence. **Includes a hard viability gate that refuses to produce a campus blueprint for a facility whose length of stay is too short to support one** — which is what makes it credible to corrections professionals.

### BEARING — three-axis classification engine
Enter a case profile; the three axes resolve independently and visibly. Surfaces the counterintuitive evidence when the profile would have triggered an offense-label exclusion, and carries the constitutional floor on screen. **This is the app that wins the argument** — it makes the separation framework concrete instead of theoretical.

### PARCHMENT — portable credential ledger
Hash-sealed, transfer-proof credential record. Each entry's seal includes the one before it, so an altered or removed entry breaks every seal that follows. Ships with an employer verification view in which offense, sentence, facility, risk score, and clinical record are **structurally absent from the payload** rather than permission-gated — there is no field to leak and nothing to subpoena.

**This is the one to build first.** A campus that cannot prove what it produced cannot be funded past its first budget cycle. And it requires no statute to change — a nonprofit can build and operate it starting now.

### CHALKLINE — 50-state readiness report card
Fifty states and DC graded on six indicators: PEP approvals, Workforce Pell adoption, registered apprenticeship in corrections, occupational licensing barriers, Medicaid reentry waiver status, and earned-time structure. Built on the PLUMBLINE pattern — every cell states whether it has been verified, no jurisdiction is graded until at least four of six indicators have been, and the gaps are displayed rather than filled in.

**Research pass 1 status: 35 cells verified, 1 of 51 jurisdictions gradeable, 220 cells still open.** That is the honest state of the field — nobody has assembled this — and the exportable research queue is what closing it looks like. It is also delegatable work, which means it can be handed to volunteers or interns without handing them judgment calls.

---

---

# Part VII — What is licensable, and to whom

Ownership first, because it governs everything below: **Born Between 2 Generals, LLC owns the entire suite.** Nonprofits — including America's Future — receive licenses. A nonprofit's involvement in a project never implies nonprofit ownership of the tooling, the data layer, or the brand. There is no exception to this in the portfolio and there is no exception here.

So the question is never "who owns this." It is "who pays for a license, and who never does."

## Who never pays

Two hard lines, and they belong in the EULA in writing rather than in a policy that a future licensing conversation can quietly erode:

**FOOTING is free to families. Permanently.** A person whose sister is sick in a facility does not hit a price. Beyond being the right answer, the free family tier is the asset — it is what makes every institutional buyer below take the call, and it is what makes the aggregate data worth anything.

**PARCHMENT is free to the credential holder. Permanently.** A record of what someone earned belongs to the person who earned it. Charging them to see it would invert the entire premise of the app.

Everything else is licensable.

## The map

| App | Buyer | Model | Revenue potential |
|---|---|---|---|
| **The CORNERSTONE standard** | Facilities and systems seeking certification | Certification fee, annual recertification, auditor training and accreditation | **Highest** |
| **PARCHMENT** | Issuers (departments of corrections, colleges, apprenticeship sponsors); employers running verifications | Per-issuance fee plus per-verification API call | **High** |
| **KEYSTONE** | State corrections departments, county sheriffs, corrections consultancies, universities, architecture and engineering firms working in corrections | Per-system annual license plus paid implementation | High, slow cycle |
| **CHALKLINE** | Advocacy organizations, media, researchers, government-affairs teams, vendors selling into corrections | Annual data subscription plus API access | Modest, steady |
| **BEARING** | Corrections departments, classification bureaus, parole boards, public defender offices, policy shops | Enterprise annual license, training add-on | Modest, narrow buyer set |
| **FOOTING — institutional tier** | Legal aid organizations, prisoners' rights groups, plaintiff-side firms, protection and advocacy agencies, oversight bodies, investigative newsrooms | Seat license for the case-aggregation layer and the jurisdiction data feed | Modest, high mission value |

## Where the actual business is

**Not in the apps. In the standard.**

Lane 3 of the national strategy — owning the accreditation framework — is also the largest revenue line in the portfolio, and the two facts are the same fact. Certification is recurring, it is defensible once adopted, it compounds as more systems certify, and it requires no legislature to authorize. The apps are what make the standard credible and operable; the standard is what makes the apps unavoidable.

**Second: PARCHMENT's verification layer.** A per-transaction credential-verification service is infrastructure rather than software. Every employer running a background check on a person with a corrections credential is a call. That volume does not depend on anyone adopting a philosophy — it only depends on the record existing.

## Cautions that belong in the contract

- **Never sell BEARING as a validated actuarial risk instrument.** It is decision support and training. It has not been validated on any population, and selling it as an actuarial tool creates liability that could reach the whole portfolio. This language belongs in the EULA, not only in the footer.
- **Government procurement needs a vehicle.** State corrections departments generally cannot buy directly. Plan for a state term contract, a cooperative purchasing agreement, or a reseller. Budget the sales cycle in quarters, not weeks.
- **PARCHMENT touches education records.** Once real learner data flows, FERPA applies to the college-issued portion, and state data-protection law may reach the rest. Structure the employer verification payload so those categories are structurally absent — which is already how it is built — and get that architecture reviewed rather than assumed.
- **Pricing is an Olson question.** Set it against a comparable scan of correctional software and accreditation fees, not against intuition. Nothing in this document assigns a dollar figure on purpose.
- **The free tiers get written into the license.** Not a promise, a term.

## Accessibility as a commercial fact

Every app in the suite now ships with read-aloud, language switching, text sizing, and high contrast built in.

This is the right thing on its own, but it is also a procurement fact worth knowing: **Section 508 and state accessibility requirements disqualify a great deal of correctional software before price is ever discussed.** Arriving already compliant removes an objection that competitors have to go build for.

FOOTING ships fully translated into Spanish. Additional languages drop into the same architecture as verified translation files — the constraint is not engineering, it is that legal and medical routing guidance must be reviewed by a native speaker before it ships, for exactly the same reason the directory carries no unverified phone numbers.

---

# Part VIII — Metrics, and what failure would look like

Publish these or the program cannot be defended:

| Metric | Window | Target |
|---|---|---|
| **Medication doses delivered as prescribed** | **Monthly** | **>98%** |
| **Days from medical request to clinical assessment** | **Monthly** | **Median under 3** |
| **Independent air and water inspections passed** | **Per schedule** | **100%, published** |
| **PREA outside-audit compliance** | **Per cycle** | **Full compliance** |
| **Grievances answered within published window** | **Monthly** | **>95%** |
| Enrollment as share of eligible population | Annual | >40% by year 3 |
| Credential completion rate | Annual | >65% of enrolled |
| Employment at 12 months post-release | Rolling | >55% |
| Wage at 12 months | Rolling | Above state median entry wage for the trade |
| Recidivism at 36 months | Cohort | 13+ point reduction vs. matched non-participants |
| Institutional violence rate | Quarterly | Declining |
| Staff assault rate | Quarterly | Declining |
| Cost per completed credential | Annual | Falling as scale grows |

**The floor metrics are not targets, they are conditions.** A facility that misses them is not a campus underperforming; it is a facility that has not earned the name yet, and its program numbers should not be reported as if it had.

**State the falsification condition up front.** If credential completion stays below 40%, or employment at 12 months does not clear 45%, or recidivism does not move by 8 points at 36 months, the model has not worked as designed and should be restructured rather than defended. Saying this before you start is what distinguishes a program from an advocacy campaign — and it is what makes a skeptical budget committee trust the rest of the numbers.

---

# Appendix A — Evidence base

| Finding | Source | Confidence |
|---|---|---|
| 43% lower odds of recidivating for correctional education participants | RAND RR-266 / RR-564 meta-analysis | Verified |
| 13 percentage-point reduction in recidivism risk | RAND RR-564 | Verified |
| $4–$5 saved per $1 invested over 3 years | RAND IG-113 | Verified |
| 13% higher post-release employment | RAND RR-564 | Verified |
| 28% higher employment for vocational participants | RAND / DOJ (2014) | Verified |
| Program cost $1,400–$1,744 per participant | RAND IG-113 | Verified |
| Pell eligibility restored for incarcerated students July 2023 | FAFSA Simplification Act; 34 CFR 668.236 | Verified |
| Workforce Pell created, implementation developing | H.R. 1, 119th Congress (2025) | Verified |
| ~68 offense categories excluded from FSA earned-time credits | 18 U.S.C. § 3632(d)(4)(D) | Verified |
| California Model / San Quentin $239M education center | CDCR; Governor's office (Feb 2026) | Verified |
| Homicide releasees among lowest rearrest rates | BJS recidivism series | Verified |
| Future-harm environmental exposure can violate the Eighth Amendment | *Helling v. McKinney*, 509 U.S. 25 (1993) | Verified |
| Deliberate indifference to serious medical needs | *Estelle v. Gamble*, 429 U.S. 97 (1976) | Verified |
| Failure-to-protect standard | *Farmer v. Brennan*, 511 U.S. 825 (1994) | Verified |
| Exhaustion of administrative remedies required | PLRA, 42 U.S.C. § 1997e(a) | Verified |

Staffing ratios, clinical cost estimates, and per-seat funding draws used in KEYSTONE are **planning assumptions, not findings**. Replace them with state figures before anything goes to a budget office.

---

# Appendix B — Open gates

Standing portfolio requirements before public use:

- **R. Emmett McAuliffe** — IP and trademark clearance on the names CORNERSTONE, FOOTING, KEYSTONE, BEARING, PARCHMENT, and CHALKLINE
- **Bill Olson** — lease, license, and EULA terms for nonprofit and public-agency deployment
- **CHALKLINE research pass 2** — 220 open cells; queue export is ready to delegate
- **FOOTING contact verification** — 561-cell jurisdiction layer built for all 51; 322 agency names carried, 239 cells open, and every populated cell still needs its contact details verified and dated before local release. Queue ships as CSV.
- **Health-department jurisdiction survey** — whether each state exempts correctional facilities from routine public health inspection is unresearched in all 51 and is the highest-value single row in the queue; it is the evidence base for floor component 1 of the model bill
- **Facility seeding** — the women's facility at the center of the current mold matter needs to be named and seeded in FOOTING
- **State-specific legal review** — the model bill package will need state counsel in each anchor jurisdiction

---

*CORNERSTONE framework, version 1.0. Born Between 2 Generals, LLC. Decision support and policy design; not legal advice, not a clinical instrument, and not a validated actuarial risk tool.*