Exploring Delay-based PUFs for Energy-Efficient Low-Overhead Security of Wearable Devices
2026-07-17T07:23:37Z•27ad148259803c7914181854da44a8959f2dbf98775977b63757da18e22abb2b
AI-environment artifactsIoT securityLLM securityMITRE ATT&CKPUFQubes OSadversary emulationagent safetyapp privacyautomated red teamingbenchmarksdata-safety labelsdifferential privacyintrusion detectionlattice cryptographylog analysismemory poisoningpassive prompt injectionprivacy measurementprompt injectionring signaturessandbox evasionvulnerability disclosurewearableszCDP
What happened
This feed collects recent security- and privacy-focused research (July 17, 2026) spanning vulnerabilities in LLM-based systems, IoT/wearable hardware, privacy measurement, sandbox evasion, automated red-teaming, and cryptography. Key findings include: LogInject — high-success (up to 88.2%) passive prompt-injection attacks against LLM-based log analysis and a Context Stitching technique to evade stateless filters; MemPoison and companion work showing persistent-memory/agent-memory prompt-injection attacks and structural blind spots that defeat simple write-time filters; AI-environment artifacts
Why it matters
A reviewed impact interpretation has not been published for this record.
Evidence and limitations
- Source ID
- arxiv_cs_cr
- Record identifier
- 27ad148259803c7914181854da44a8959f2dbf98775977b63757da18e22abb2b
- Enrichment time
- 2026-07-17T07:23:37Z
- 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.