Property-Level Reconstructability of Agent Decisions: An Anchor-Level Pilot Across Vendor SDK Adapter Regimes

TL;DR

This study evaluates how well different vendor SDK regimes allow reconstruction of agent decisions at the property level, revealing variability and gaps in reconstructability across regimes.

Contribution

It introduces a schema for classifying property-level reconstructability of agent decisions and applies it across six vendor SDK regimes, highlighting regime-dependent gaps.

Findings

Reconstructability varies significantly between regimes.

Strict governance completeness ranges from 42.9% to 85.7%.

Identifies a regime-independent gap in reasoning trace reconstruction.

Abstract

Agentic AI failures need post-hoc reconstruction: what the agent did, on whose authority, against which policy, and from what reasoning. Cross-regime feasibility remains unmeasured under one property-level schema. We apply the Decision Trace Reconstructor unmodified to pinned worked-example anchors from six public vendor SDK regimes spanning cloud-agent, observability, tool-use, telemetry, and protocol traces, plus two comparator columns. Each Decision Event Schema (DES) property is classified as fully fillable, partially fillable, structurally unfillable, or opaque. Per-property reconstructability of an agent decision already varies between regimes at this anchor scale. Strict-governance-completeness separates into three tiers ranging from 42.9% to 85.7%, yielding one regime-independent gap (reasoning trace), four regime-dependent gaps, and one Mixed property; the pilot is single-annotator, one anchor per cell, descriptive, with outputs checksum-verifiable from a deposited reproducibility package.