> 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/8.-security-and-enforcement/8.8-continuous-admissibility.md).

# 8.8 Continuous Admissibility

Some AI-triggered actions are effectively instantaneous.

Others may continue over time.

Examples may include:

* long-running infrastructure changes,
* multi-stage financial workflows,
* autonomous remediation,
* distributed agent tasks,
* deployment operations,
* or physical AI activity.

For these systems, a one-time admission check may be insufficient.

The hardened security model therefore defines **Continuous Admissibility**.

### Core Invariant

For a governed execution interval:

```
t ∈ [T₀, T₁]
```

the applicable authority condition should remain valid throughout the period in which continued authority is required.

Conceptually:

```
∀t ∈ [T₀, T₁]:
AuthorityState(t) = ACTIVE
```

otherwise:

```
Execution halts
```

### Why Initial Admission Is Not Enough

Consider:

```
T0:
Action admitted

T1:
Execution begins

T2:
Authority revoked

T3:
Execution continues

T4:
Irreversible consequence occurs
```

A purely pre-execution system could have been correct at T0 while still producing illegitimate consequence after T2.

Continuous admissibility addresses this gap.

### Continuous Does Not Necessarily Mean Every CPU Cycle

The architectural requirement is that the enforcement frequency must be appropriate to the execution and consequence profile.

Possible mechanisms may include:

* continuous revocation channel,
* periodic revalidation,
* event-driven revocation updates,
* short-lived Permits,
* per-stage Permit renewal,
* or mandatory re-admission at defined boundaries.

The exact mechanism is deployment-specific.

### Material Mutation

Continuous admissibility also interacts with trajectory governance.

If execution or agent planning undergoes:

```
Material Mutation
```

then:

```
Re-Admission
```

may be required even if the Authority Object itself remains active.

Thus continuous governance may monitor both:

```
Authority State
```

and:

```
Admission-Relevant Trajectory State
```

### Interruptibility

Continuous admissibility has value only if the downstream system can act on the changed governance state.

For long-running execution, the architecture therefore depends on some form of interruptible execution boundary.

### Specification vs Validation

Continuous admissibility is formally present in the Authority Layer Security Specification and CP / EP separation model.

However, the current six core execution scenarios primarily demonstrate:

* revocation state enforcement before a later execution attempt,
* Permit verification,
* replay rejection,
* scope enforcement,
* and related execution-boundary properties.

The current public evidence should therefore not be interpreted as comprehensive validation of continuous mid-execution revocation across arbitrary long-running or distributed production systems.

Continuous admissibility is:

```
Specified
```

and represented architecturally.

Its full production validation remains profile-dependent.