# NOTICE

superred-claim-agentdojo

This package's own source code is original work, released under the MIT
License (see LICENSE), Copyright (c) 2026 Simon Sure.

It bundles no third-party code and no third-party datasets. The following
upstream project is a RUNTIME DEPENDENCY (installed separately from PyPI,
not redistributed inside this wheel) and is credited here because Layer 1
of this SecurityClaim is a faithful adapter of the AgentDojo benchmark.

## AgentDojo

Layer 1 runs and scores the original AgentDojo injection/user tasks by
importing the AgentDojo library at runtime (task base classes, task-suite
loader, and its security/utility trace predicates); it also references
AgentDojo's task identifiers. No AgentDojo source code or data is copied
into this package.

- Project: AgentDojo (Debenedetti et al., NeurIPS 2024 Datasets & Benchmarks)
- Repository: https://github.com/ethz-spylab/agentdojo
- License: MIT
- Copyright (c) 2024 Edoardo Debenedetti, Jie Zhang, Mislav Balunovic,
  Luca Beurer-Kellner, Marc Fischer, and Florian Tramèr

## Citation

When reporting numbers produced with this module, cite Debenedetti et al.,
"AgentDojo: A Dynamic Environment to Evaluate Prompt Injection Attacks and
Defenses for LLM Agents" (2024).
