ADeptS-Bench: Measuring the Trustworthiness of Computer Use Agents Across Devices

arXiv 2608.2620427596f7270fbd830794631dd71dc5f915e739a948d8e34e30b463289dbc2bd6d
AI agent safetyAI model integrityCHERICTF evaluationCWEIoMTLLM code generationRTL securityTEETOCTOUXRP Ledgercomputer-use-agent securityconsensus robustnessenclave isolationhardware securityintegrity attacksoffensive securityprompt ambiguityremote attestationreplay attackstemporal memory safetytrace provenanceuse-after-freeuse-after-returnweight obfuscation

Paper metadata

arXiv ID
2608.26204
Version
Not specified by this published record
Category
Computer Science — Cryptography and Security (cs.CR)

The PDF link points to arxiv.org. Baitaphish does not expose a private stored PDF.

Evidence and limitations

Source ID
arxiv_cs_cr
Record identifier
27596f7270fbd830794631dd71dc5f915e739a948d8e34e30b463289dbc2bd6d
Enrichment time
2026-08-28T07:23:27Z
AI-assisted enrichment
Yes

This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.