MMODELYST
Papers/RedAct: Redacting Agent Capability Traces for Procedural Skill Protection
PAP

RedAct: Redacting Agent Capability Traces for Procedural Skill Protection

Jun 10, 2026

arXiv
Abstract

Users rely on execution traces to observe agent behavior, diagnose failures, and ensure accountability. These traces contain rich procedural detail, including tool invocations, intermediate decisions, and error-recovery logic. Yet this detail can expose private procedural skills, allowing downstream methods to recover key formulas, thresholds, and strategies without access to model weights or skill files. To quantify this risk and evaluate protection, we construct CapTraceBench, a benchmark of 75 specialized long-horizon tasks and 154 curated skills across seven domains. We also introduce RedAct https://github.com/XuShuwenn/RedAct, a protected trace release framework that localizes protected key information, rewrites traces while preserving verifier-critical evidence, and embeds behavioral watermarks for downstream provenance analysis. Across representative trace reuse methods, RedAct reduces normalized skill transfer (NST) from 44.7--67.1\% on raw traces to below the no-skill baseline, while preserving audit evidence. Its standalone behavioral watermarks reach 93.6--100.0\% true detection with a false alarm rate of at most 1.9\%. These results frame public agent traces as security interfaces and show that selective redaction can reduce procedural capability leakage without removing audit evidence.

Select text to highlight · click a highlight to remove · saved in this browser only
Authors
Shuwen Xu, Zhitao He, Yi R. Fung
Your notes (browser-local)
saved
Cross-links
No linked entities.
arXiv:2606.10813