> For the complete documentation index, see [llms.txt](https://docs.lpp-minduniverse.org/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.lpp-minduniverse.org/lingua-pactum-protocol-lpp-documentation/9.-deployment-profiles/9.-deployment-profiles.md).

# 9. Deployment Profiles

LPP defines a general constitutional admissibility model.

Deployment is domain-specific.

The same canonical governance sequence applies across environments:

```
Authority
→ Intent
→ Admissibility
→ Permit
→ Execution
→ Evidence
```

What changes is the concrete meaning of:

* authority,
* scope,
* consequence,
* delegation,
* revocation,
* execution boundary,
* required verification,
* and reconstructable evidence.

The central deployment question is therefore:

> **What could this AI system actually cause to happen, and under whose legitimate authority?**

An AI agent connected only to an informational system presents a different governance problem from an agent capable of:

* transferring assets,
* changing production infrastructure,
* executing financial transactions,
* controlling industrial equipment,
* or coordinating consequential actions across several principals.

LPP does not solve these domain differences by creating a separate constitutional protocol for every industry.

Instead:

```
General LPP Constitutional Model
        ↓
Domain Governance Profile
        ↓
Domain-Specific Authority / Scope / Consequence Rules
        ↓
Execution Boundary
```

The canonical Layer 0 principles remain constant:

```
Capability
≠
Authority
```

```
Runtime Permission
≠
Constitutional Admissibility
```

```
Failed Mandatory Predicate
cannot be compensated downstream
```

```
Only an admissible action
may proceed toward execution
```

The following deployment profiles illustrate how these principles apply across representative AI execution environments.

***