masahiroid/injecagent-ja

InjecAgent-ja (v0.1.1, translation-first) English | 日本語 Summary A Japanese translation of InjecAgent (Zhan et al., ACL Findings 2024; MIT): 1,054 test cases (510 direct-harm + 544 data-stealing) x 2 settings (base, enhanced) for evaluating indirect prompt injection against tool-using LLM agents. Translated: attacker instructions, expected achievements, user instructions, the agent's Thought, and the natural-language values inside tool-response templates; the… See the full description on the dataset page: https://huggingface.co/datasets/masahiroid/injecagent-ja.

<iframe src="https://silytics.si/embed/item/e92df7cc-c639-4cb4-81c1-c345a66ba4af" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.