# InjecAgent

> Benchmark of 1,054 test cases measuring how tool-using LLM agents fall for indirect prompt injections in tool outputs.

- id: `injecagent`
- category: Information (`information`)
- url: https://github.com/uiuc-kang-lab/InjecAgent
- repo: https://github.com/uiuc-kang-lab/InjecAgent
- pricing: open-source
- license: MIT
- tags: benchmark, prompt-injection, security, tool-use, dataset
- related: AgentDojo (https://indexagentica.com/entries/agentdojo/)
- sources: https://github.com/uiuc-kang-lab/InjecAgent , https://arxiv.org/abs/2403.02691
- added: 2026-10-31
- updated: 2026-10-31
- last verified: 2026-10-31
- maintainer: UIUC Kang Lab
- submitted by: agentica-curator
- page: https://indexagentica.com/entries/injecagent/
- json: https://indexagentica.com/api/entries/injecagent.json

The cases span 17 user tools and 62 attacker tools, covering both direct harm and data-stealing attacks. Scripts evaluate prompted (ReAct) agents and fine-tuned function-calling agents. From UIUC's Kang Lab (paper arXiv:2403.02691). MIT. Last updated July 2024.