benchmark · first seen 16 Oct 2026
AgentDojo
A benchmark environment for evaluating prompt injection attacks and defenses in tool-using LLM agents.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Ignored, Not Refused: What Low Prompt-Injection Success Rates Hide in Small Tool-Using AgentsPapers · 16 Oct 2026 · Farhan Zarif, Kazi Shahid Ahmed Galib, Synthia Akter Promi, Shabnam Sharmin Tari…
- What Standard Evaluations Hide About Detectors of Hijacked LLM Agents: Rare Attacks, Unseen Tools and Benchmark TransferPapers · 30 Oct 2026 · Mahmoud Abbasi, Juan M. Núñez V., Amin Shahraki
- CapLite: Prompt-Injection Defense for Small Language Model Agents without Model-Written Security ArtifactsPapers · 12 Nov 2026 · Zimo Xu, Wenzhuo Liu, Hideki Nakayama
Latest sources
- CapLite: Prompt-Injection Defense for Small Language Model Agents without Model-Written Security ArtifactsPapers · 12 Nov 2026 · Zimo Xu, Wenzhuo Liu, Hideki Nakayama
- What Standard Evaluations Hide About Detectors of Hijacked LLM Agents: Rare Attacks, Unseen Tools and Benchmark TransferPapers · 30 Oct 2026 · Mahmoud Abbasi, Juan M. Núñez V., Amin Shahraki
- Ignored, Not Refused: What Low Prompt-Injection Success Rates Hide in Small Tool-Using AgentsPapers · 16 Oct 2026 · Farhan Zarif, Kazi Shahid Ahmed Galib, Synthia Akter Promi, Shabnam Sharmin Tari…