Publications

Active Discovery of Hidden Business Cascades under Rule Drift

Hidden business dependencies and cascades as operational rules change.

Time-Consistent Evidence Assembly across Documents and Transactional Data for AI Agents

Time-consistent evidence drawn from documents and transactional records.

Online Context Scheduling under Cache Expiry and Human Approval Delays

Context scheduling when cached information expires and human approvals take time.

Preserving Business Semantics in Multilingual Tool Workflows: Contract Checks and Model-Failure Diagnosis

Business-semantic consistency, contract checks, and failure diagnosis across languages.

Dependency-Complete Tool Discovery in Evolving Enterprise Tool Catalogs

Tool discovery that accounts for dependencies in changing enterprise catalogs.

Incremental Repair of Agent Memory and Plans after Business Evidence Revisions

Incremental updates to agent memory and plans after business evidence changes.

Clarification Planning for AI Agents under Shared Question Budgets

Clarification planning when agents share a limited question budget.

Repairing AI Agent Task Graphs: Audit-Guided Recomputations versus Full-Context Rewriting

Task-graph repair through audit-guided recomputation and full-context rewriting.

Evaluating Lifecycle Interactions in AI Agents across Memory, Approval, Execution, and Recovery

Interactions among memory, approval, execution, and recovery across an agent lifecycle.

Evidence-Guided Recovery for AI Agents with Ambiguous Tool Outcomes and Partial Execution

Evidence-guided recovery when tool outcomes are ambiguous or execution is partial.

Verifying Text-to-SQL Queries under Changing Business Semantics

Verification of generated SQL as business meanings and rules change.

Change Detection for Task Resumption in AI Agents: An Empirical Study

Change detection when an AI agent resumes a paused task.

Authorization-Preserving Context Lifecycles for Long-Horizon AI Agents

Authorization boundaries throughout long-horizon agent context lifecycles.

Authorization for Self-Modifying AI Agent Populations: Conserving Authority across Replacement, Forking, and Rollback

Authority conservation as agent populations replace, fork, and roll back software.

Fault-Tolerant Budget Conservation in Distributed Multi-Agent Delegation

Budget conservation and fault tolerance in distributed multi-agent delegation.

Language-Conditioned Token and Reasoning Efficiency in Large Language Models: A Paired Cross-Lingual Study Protocol

A paired cross-lingual study protocol for token use and reasoning efficiency.

Runtime Authorization of Self-Generated Subgoals in Long-Horizon Tool-Using AI Agents

Runtime authorization of subgoals generated by long-horizon tool-using agents.

Authorization Revocation for Long-Running AI Agents: Root-Scoped Quiescence under Delegation and Asynchronous Execution

A root-scoped revocation protocol that fences delegated and asynchronous old-root paths while preserving work backed by current, independently sufficient authority.

AcquireBound: Runtime Authorization for Resources Acquired by AI Agents

A provenance-bounded runtime that quarantines agent-acquired resources, resolves their real capabilities, and activates only authority within a current relational policy envelope.

Versioned Transitive Dependency-Closure Binding and Operation-Time Effect Governance for Agent Skills: ClosureBound

A reference monitor that binds authorization to the exact transitive dependency closure of an agent skill and revalidates that closure when a durable effect is produced.

International Migration, Housing Demand Measurement, and Vancouver New-House Price Forecasting

A reproducible measure of migration-associated occupied-dwelling use, evaluated for incremental forecasting value against Vancouver new-house price benchmarks.

Explanation-Bound Tool Execution for AI Agents: Server-Verified Action Claims Without Trusting Model Rationales

A claim-carrying mediation layer that converts decision-relevant rationale content into typed action claims and verifies them against current server-held facts before governed execution.

Intent-Governed Tool Authorization for AI Agents

A server-side authorization layer that binds tool visibility and proposed effects to the user’s current expressed intent—narrowing, never expanding, existing authority.

OpenPort Protocol: A Security Governance Specification for AI Agent Tool Access

A governance-first specification for exposing application tools to AI agents through server-side authorization, reviewable writes, predictable failure semantics, and audit.