Research
Dive into Claude Code: The Design Space of Today’s and Future AI Agent Systems
Publication date unknown · Discovered arxiv.org
Page published
Publication date unknown · First observed: 25 Oct 2026
Coverage timeline
Single-source research — one report is available.
Why it matters
Claude Code’s permission architecture and the contrasting designs of OpenClaw and Hermes Agent help defenders evaluate the security boundaries protecting systems accessible to autonomous agents.
Dive into Claude Code analyzes publicly available source code and compares the permission systems, execution architectures, and extensibility mechanisms of Claude Code, OpenClaw, and Hermes Agent. The study contrasts per-action controls with perimeter-level access controls and cross-surface approvals. A linked Adversa AI report describes a Claude Code weakness that bypasses configured deny rules when shell commands contain more than 50 subcommands.