- Rating
- 2
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
This paper studies how trajectory poisoning can subvert the safety harness of self-evolving LLM agents. It proposes TRAJPOISON, an attack framework that presents attacker-chosen harness edits as legitimate safety repairs, refines poisoned trajectories using diagnostic feedback from a surrogate evolution engine, and adds supporting trajectories to help the edits pass validation. The attack can weaken safety constraints or induce excessive restrictions that reduce utility. The paper also introduces EvoGuard, a defense that examines the evidence supporting candidate updates, replays tasks to check whether safety-relevant observations reproduce, and rejects updates that expand permitted actions on harmful probes in a private invariant suite. Experiments show that TRAJPOISON degrades safety or utility, with these effects transferring to a held-out benchmark and other agent backbones. In the primary evaluation setting, EvoGuard substantially mitigates the attacks while largely preserving the benefits of benign evolution.
Critical strengths
-
The threat model is plausible, and the problem addresses a real-world problem.
-
The evaluation covers both poisoning modes and checks transfer across benchmarks and agent backbones. These settings fit the claims about attack effectiveness.
Critical weaknesses
-
The proposed methods are insufficiently validated. This work promised to compare TRAJPOISON with existing poisoning methods, but it doesn't actually test against poisoning baselines. The promised ablation in the Reproducibility Statement is also not provided. Moreover, it does not directly evaluate the evidential asymmetry underlying EVOGUARD.
-
The claim about persistence lacks experimental support. The abstract and introduction claim the attack effects last for 20 rounds after removal, but the experiment applies poisoning in every round.
Questions for authors
-
How is C(∆) calculated, and where does the threshold come from?
-
Besides safety-invariant verification, what results show that evidence dispersion and reproducibility score can distinguish benign patches from poisoned patches and improve EVOGUARD's effectiveness?
-
Which experimental results support the 96.8% reduction in attack success rate claimed in the abstract and the 96.8% reduction in harmful rate claimed in the introduction?
-
What experimental results support the claimed persistence for 20 rounds after the poison is purged?
-
How does TRAJPOISON perform compared with existing poisoning attacks, and what experimental results show the contribution of each stage?