Volume V — Stable Authority Boundary™ and Agentic Systems

The Authority Problem in AI, Automation, and Distributed Systems

BookPublication-ReadyAvailable for ReviewRights AvailablePublic Preview

This work is publication-ready and available for publisher, licensing, or strategic review unless and until publication rights, exclusive review rights, or another contractual arrangement is accepted.

Complete manuscript available for publisher, licensing, or strategic review. This page provides a controlled public sample only.

Discovery metadata

Keywords

Stable Authority BoundarySABAURORAARCautonomous systemsauthority boundaryrefusalsilencerecoveryevidencereceiptsauditdistributed systemsresilienceAgentic Systems The Authority ProblemDistributed Systemsauthoritycapabilityboundaryfailureagenticshouldautonomousmission
PUBLIC SAMPLE

Sample Chapter

Chapter 1 - The Authority Problem in Agentic Systems

For most of the history of computing, authority was relatively easy to understand.

Software executed instructions. Operators issued commands. Managers approved actions. Engineers designed systems. Authority remained largely attached to identifiable people and organizational structures. The machine performed work, but responsibility for the decision to perform that work remained visible.

Agentic systems challenge that simplicity.

Modern systems are increasingly capable of selecting actions, adapting plans, coordinating resources, invoking tools, supervising subordinate processes, and operating for extended periods with limited human intervention. They do not merely execute commands. They participate in deciding how commands should be carried out.

As a result, a fundamental question emerges:

Who—or what—is exercising authority?

The answer is often less clear than designers would like to admit.

Many autonomous systems are built around a capability model. Engineers identify functions the system must perform, define operating constraints, establish objectives, and then optimize the system's ability to achieve those objectives. The architecture focuses on performance, efficiency, reliability, safety, accuracy, resilience, and adaptability.

These are all necessary concerns.

Yet they leave an important question unanswered.

What makes a particular action legitimate?

The existence of capability does not automatically create authority.

A navigation system may be capable of rerouting a convoy.

An autonomous aircraft may be capable of selecting an alternate flight path.

An industrial controller may be capable of shutting down a production line.

An artificial intelligence system may be capable of approving, rejecting, escalating, recommending, coordinating, or allocating resources.

Capability answers the question:

Can the system perform the action?

Authority answers a different question:

Is the system legitimately allowed to perform the action?

The distinction appears obvious when stated directly.

Unfortunately, many architectures behave as though the distinction does not exist.

When systems are given objectives, they frequently infer that achieving the objective justifies the actions required to achieve it. When communication is degraded, systems often continue operating based on prior assumptions. When authority becomes ambiguous, systems frequently default toward continued execution rather than constrained behavior. When uncertainty increases, many architectures encourage action instead of restraint.

The result is an implicit expansion of authority.

No formal transfer occurs.

No authorized official signs a document.

No command structure deliberately grants additional power.

Yet the system begins acting as though authority has expanded because capability remains available.

This phenomenon appears repeatedly across technical domains.

Distributed systems continue operating under stale assumptions.

Autonomous vehicles make decisions outside originally intended operating conditions.

Software agents invoke tools beyond the

scope envisioned by their operators.

Machine-learning systems produce recommendations that gradually become de facto decisions.

Organizations discover that operational authority has migrated from people to processes without anyone consciously approving the transfer.

In each case, the failure is not necessarily malicious.

The failure is architectural.

The system was designed to answer whether it could act but was never required to continuously evaluate whether it should act.

This is the Authority Problem.

The Authority Problem exists whenever capability remains available after legitimate authority has become uncertain, degraded, exceeded, transferred, withdrawn, expired, or was never properly established in the first place.

Most systems possess mechanisms for execution.

Far fewer possess mechanisms for legitimacy.

This imbalance becomes increasingly dangerous as systems become more autonomous.

The more capable a system becomes, the more opportunities it has to exceed the authority originally intended for it.

A calculator presents little risk.

A distributed autonomous architecture capable of coordinating hundreds of systems presents a very different challenge.

As capability expands, authority boundaries become more important, not less.

The traditional response has been to add policies, procedures, permissions, governance documents, review boards, compliance frameworks, and human oversight. These measures provide value, but they often exist outside the operating logic of the system itself.

The system may be governed by authority without understanding authority.

This volume argues that authority must become an executable architectural concern.

Authority cannot remain merely an organizational concept documented outside the system. It must become part of the system's operating model.

Before we can design authority-bounded autonomous systems, we must first understand why current architectures repeatedly drift toward capability-centered behavior.

The remainder of this chapter examines the mechanisms that create that drift and why increasingly autonomous systems make the problem impossible to ignore.

Capability Drift and Implicit Authority Expansion

Most authority failures do not begin with deliberate abuse.

They begin with success.

A system is given a task. The system performs the task successfully. Confidence grows. Additional responsibilities are assigned. Additional capabilities are added. Additional automation is introduced. More decisions are delegated. More actions become automatic.

Each step appears reasonable.

Each step appears justified.

Each step appears beneficial.

Over time, however, a subtle transition occurs.

The system is no longer being trusted because its authority has been carefully examined. The system is being trusted because it has been successful.

Capability begins replacing authority as the basis for decision-making.

This transition is rarely documented.

No organization typically announces:

"We are expanding authority without review."

Instead, authority expansion occurs indirectly through operational convenience.

A system that consistently performs one task successfully is allowed to perform a second.

A system that assists decision-making becomes responsible for filtering available options.

A system that filters options begins recommending actions.

Recommendations become defaults.

Defaults become expectations.

Expectations become operational practice.

Operational practice becomes assumed authority.

At no point does a formal authority transfer necessarily occur.

Yet authority has effectively expanded.

The system now influences outcomes beyond its original

scope…

Table of Contents

Complete contents extracted from the current manuscript source.

Complete contents — 251 entries · through page 367
Preface11
Capability Is Not Authority11
Part I — The Core Failure15
Chapter 1 - The Authority Problem in Agentic Systems19
Capability Drift and Implicit Authority Expansion23
The Assumption of Objective Legitimacy28
What This Chapter Established33
Key Principles34
Chapter 235
Chapter 2 - Authority vs Capability37
Capability Is Not Permission40
A Simple Authority Test43
Authority as a Runtime Condition44
Authority may decrease even while capability remains unchanged.48
Sources of Authority50
Authority Flow and Authority Boundaries53
Authority should be easier to lose than to gain.56
Authority Under Uncertainty56
The Cost of Confusing Access with Authority60
Why Authority Must Be Explicit63
Authority as an Architectural Primitive67
Chapter Summary69
Chapter 3 — Delegated Capability73
The problem is unbounded capability.74
Authority must be granted.76
Not every authority should be delegated.79
Capability can make a system useful.81
Delegation is a chain.82
Part II — SAB Applied to Agentic Architecture85
Chapter 4 — Stable Authority Boundary™ in Autonomous Execution89
Access is not authority.90
A Stable Authority Boundary™ should make boundary status visible.93
Silence must be governed.97
That is governance.99
That is precisely why the boundary risk is so serious.100
That verification is the point of SAB.103
This is where SAB imposes discipline.104
The agent should stop in a governed way.107
A capable system can perform a task.109
This is the operational maturity SAB provides.111
Chapter 5 — Authority Transfer Without Authority Expansion113
A Stable Authority Boundary™ does not prevent authority from moving.113
The boundary travels with the task.116
The transfer is always conditional.121
This is the balance SAB requires.123
This distinction is central to SAB.125
Need is not authority.127
Chapter 6 — Executable Refusal131
An architecture prevents the unauthorized act from occurring.133
The ease of execution does not reduce the authority required.135
Executable refusal blocks that failure mode.136
A SAB-governed agent should behave differently.138
A system that refuses too early may become useless.140
Readiness is not authorization.142
SAB changes the success condition.145
Part III — AURORA™ and Capability Directives™151
Chapter 7 — AURORA™ Capability Directives™ (ACD)163
An AURORA™ Capability Directive™ is different.164
An ACD is not merely a better instruction.168
This makes the ACD useful across many scales.171
This is why ACDs are central to AURORA™.172
An ACD reduces that uncertainty.174
It can also create mistakes.176
This is what gives the ACD practical force.177
An ACD is not just a command wrapper.178
This does not eliminate human support judgment.180
An ACD keeps the system inside the lane it was given.182
Chapter 8 — Mission Package Containment185
A mission package is not an open invitation to act.185
Containment prevents that drift.186
This is where containment supports evidence discipline.188
Containment reduces guesswork.190
Termination is not failure.191
Chapter 9 — Capability Directives™ vs. Prompts, Policies, and Rules195
AURORA™ Capability Directives™ are not prompts.195
Policies have a different limitation.197
ACDs offer a way to think differently about that problem.199
That is what an ACD provides.204
Part IV — Distributed Authority and Failure207
Chapter1
0 — Distributed Authority Under Node Failure213
Distributed systems fail in ways centralized systems do not.213
A Stable Authority Boundary™ cannot allow that.215
That is the SAB discipline.216
The system should leave evidence of those decisions.218
1 — Acting Master, Worker Refusal, and Authority Contraction221
It is boundary preservation.223
SAB therefore treats failure as a boundary-sensitive condition.226
This principle protects the mission.228
2 — Degraded Coordination and Bounded Continuation231
A degraded system should not automatically stop.231
A degraded output should not pretend to be a normal output.233
SAB does not reject continuity.235
Receipts prevent that confusion.236
Part V — Receipts, Accountability, and Review239
3 — Receipts and Accountability for Autonomous Systems245
Autonomous systems create an accountability problem because they act through sequences.245
SAB receipts must answer a deeper question:247
Receipts make the chain visible again.250
A receipt-based system changes that posture.252
A receipt-based system is different.254
4 — Human Review Boundaries255
This changes the role of human review.256
That boundary protects both sides.258
That is the human review boundary.260
5 — Failure Review: When the System Did the Right Thing by Not Acting263
Failure review must be able to tell the difference.264
This is the point of failure review under SAB.266
A SAB-governed system must do better than both.269
Part VI — Implementation Framework271
6 — Designing SAB-Compatible Agentic Systems277
An agentic system is not made safe by intelligence alone.277
That is why design matters.279
The cost of ignoring this design discipline is not merely technical risk.281
SAB reduces that uncertainty.283
7 — Testing Authority-Bounded Autonomy285
An authority-bounded system must be tested differently from an ordinary application.285
This is a different testing discipline.286
Refusal testing is equally important.287
Testing should prove that the room has walls.289
8 — Deployment Patterns293
The Stable Authority Boundary™ is not limited to artificial intelligence.293
The pattern changes.294
That is the broader value of SAB.297
SAB does not make these systems less useful.298
Closing — The Boundary You Do Not See299
9 — The Future of Legitimate Autonomy305
The future will not contain fewer autonomous systems.305
Capability grows, but accountability thins.306
SAB offers a different future.307
Legitimacy requires more.309
Appendices311
Appendix A — SAB Agentic Glossary313
Accountable Authority313
Acting Master313
Agentic System313
Authority Boundary314
Authority Condition314
Authority Contraction314
Authority Drift315
Authority Residue315
AURORA™ Capability Directive™ / ACD315
Boundary Event316
Bounded Autonomy316
Bounded Continuation316
Capability316
Capability Registry317
Command Authority317
Degraded Authority317
Degraded Coordination318
Delegated Capability318
Evidence Boundary318
Evidence Discipline318
Executable Refusal319
Governed Degradation319
Human Review Boundary319
Mission Package319
Mission Package Containment320
Node320
Partial Authority320
Receipt320
Receipt Ledger321
Refusal321
Safe State321
Stable Authority Boundary™ / SAB322
Tool Access322
Worker Node322
Appendix B — AURORA™ Capability Directive™ Template323
ACD Reference Header323
Purpose and Authority324
Actor or Component Assignment324
Capability Definition325
Data and Evidence Boundary326
Operating Conditions327
Boundary Behavior327
Support and Mitigation328
Receipt Requirements329
Human Review329
Directive Closure330
Appendix C — Mission Package Containment Checklist331
Mission Identity331
Mission Objective332
Authorized Capabilities332
Capability Limits332
Data Boundary333
Evidence Use333
Data Degradation334
Authority Limits334
Human Review Boundary334
Escalation Path335
Refusal Path335
Safe-State Conditions335
Distributed Operation336
Acting Role Limits336
Termination Conditions336
Receipt Requirements336
Reviewability337
Containment Summary337
Appendix D — Executable Refusal Test Cases339
Test Case 1 — Draft Versus Release339
Test Case 2 — Candidate Finding Versus Official Determination339
Test Case 3 — Available Source Outside Evidence Boundary340
Unauthorized340
Test Case 5 — Urgent Request Without Authority341
Test Case 6 — Degraded Evidence341
Test Case 7 — Expired Mission Package342
Test Case 8 — Human Review Boundary Reached342
Test Case 9 — Acting Master Overreach343
Test Case1
0 — Support Restoration Without Authority343
1 — Dashboard Update as Operational Record344
2 — Medical Support Without Clinical Judgment344
3 — Financial Approval Boundary345
4 — Public Communication Boundary345
5 — Refusal Explanation Quality345
6 — Refusal Too Early346
7 — Refusal Too Late346
8 — Boundary-Aware Recovery347
Appendix E — Receipt Schema for Authority-Bounded Systems349
Receipt Header349
Authority Context350
Mission and Directive Context350
Actor Context351
Request Details352
Evidence Context352
System Behavior353
Outcome353
Integrity and Review354
Retention and Access354
Receipt Sufficiency Test355
Appendix F — Distributed Node Failure Scenarios357
Scenario 1 — Master Node Failure With Worker Continuation357
Scenario 2 — Acting Master Assumes Limited Coordination357
Scenario 3 — Worker Refuses Acting-Master Instruction358
Scenario 4 — Communications Degrade but Do Not Fail Completely359
Scenario 5 — Local Node Retains Capability but Loses Authority Refresh359
Scenario 6 — Source Degradation During Distributed Operation360
Scenario 7 — Split Authority View360
Scenario 8 — Delayed Handoff Receipt361
Scenario 9 — Support Intervention During Degraded Authority361
Scenario1
0 — Node Produces Candidate Output Under Degraded State362
1 — Local Safety Preservation362
2 — Expired Mission Package in Distributed State363
3 — Reconnection After Divergent Operation363
4 — Acting Master Attempts External Communication364
5 — Worker Receives Conflicting Directives364
6 — Degraded Node Attempts Irreversible Action365
7 — Silent Continuation After Failure365
8 — Failure Review Finds Correct Non-Action366
Appendix Summary367