A Contract-Centered Architecture for Scalable and Manageable Agentic Runtimes
cs.AI, cs.MA, cs.SE
Submitted: 2026-08-27
Updated: 2026-09-12
Comments: 52 pages, 8 figures; revised data-use contracts with an executable reference model and finite synthetic conformance artifact; updated related work and experimental decision rules. Runtime scaling remains empirically untested
License: http://creativecommons.org/licenses/by/4.0/
The gist: Enterprise agentic systems must coordinate changing capabilities, execution capacity, and independently governed data.
Terminology
Abstract
Enterprise agentic systems must coordinate changing capabilities, execution capacity, and independently governed data. We define Skill, Harness, Scaffold, and an external data substrate as responsibility contracts. The central hypothesis, cost-aware capability-capacity separability, asks whether compatible capacity changes preserve semantic outcomes while capability changes preserve the capacity-response relationship within declared margins and enforcement budgets. We operationalize the data boundary through a source-oriented Data Wiki, an output-oriented Theme Wiki, and a versioned Intermediate Relation. Executable 5W1H+Which predicates bind source identity, validity, authorization, semantics, operations, relations, and evidence requirements. Request-bound tickets add execution-time revalidation and typed rejection. A conditional soundness argument states the required trust and atomicity assumptions; dependency invalidation makes change propagation explicit. A single-process reference model agrees with a declared specification oracle on all 1,024 combinations in a finite synthetic fault domain and passes five lifecycle checks. These are conformance results, not measurements of retrieval quality, production safety, or scaling. We specify a held-out data study and a cluster-period crossover with distinct supported, falsified, conditional-engineering, and inconclusive verdicts. The core separability hypothesis remains empirically untested.
Sources
- A Policy-Driven Runtime Layer for Agentic LLM Serving
- Tool Forge: A Validation-Carrying Toolchain for Governed Agentic Execution
- AEVAL: From Anecdotal to Deterministic Testing for Agentic Skill Workflows
- Agent Skill Security: Threat Models, Attacks, Defenses, and Evaluation
- DeltaBox: Scaling Stateful AI Agents with Millisecond-Level Sandbox Checkpoint/Rollback
- Skillware: A Software Ontology and Engineering Lifecycle for Persistent Behavioral Artifacts
- From Model Scaling to System Scaling: Scaling the Harness in Agentic AI
- Is Progressive Disclosure All You Need for Long-Context Agents?
- Scoped Verification for Reliable Long-Horizon Agentic Context Evolution under Distribution Shift
- MemoHarness: Agent Harnesses That Learn from Experience
- When Summaries Distort Decisions: Information Fidelity in LLM-Compressed Financial Analysis
- Workflow-Localized Mechanism Learning: Attribution-Guided Repair and Knowledge Reuse for Structured Agent Skills
- Agentic Coding in the Wild: Characterizing GitHub Copilot Traces at Production Scale
- HarnessBank: Semantic Gene-Bank Search with Gated Verification for Agent-Harness Self-Evolution
- CoAgent: Concurrency Control for Multi-Agent Systems
- Code as Agent Harness
- Falsifiable Release Gates for Self-Improving Systems: Standing Invariants at Scale
- A Five-Plane Reference Architecture for Runtime Governance of Production AI Agents
- Beyond Top-K: Replacing Black-Box Retrieval with Interpretable Agentic Operations
- BM25 Wins at Scale: A Scaling Study of Retrieval-Augmented Generation Paradigms
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Agents' Last Exam
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems