Metraipolis › archive › dc_40bb5d439f04

Can Benign Skills Induce Misaligned Behavior in LLM Agents? — arXiv:2611.04299

Source: arXiv (en)
Original: https://arxiv.org/abs/2611.04299
License (as archived): CC0-1.0
Archived: 2026-11-04T21:49:40.378Z
SHA-256 of the archived text: 706ddfae39d4935af7f555957f92c432fa51d4b1d0f972ebc83fa85962090d56
Stable URL of this page: https://metraipolis.com/documento/dc_40bb5d439f04

How to cite

Can Benign Skills Induce Misaligned Behavior in LLM Agents? — arXiv:2611.04299 — arXiv (en), https://arxiv.org/abs/2611.04299 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_40bb5d439f04 (Metraipolis archive dc_40bb5d439f04, sha256 706ddfae39d4935af7f555957f92c432fa51d4b1d0f972ebc83fa85962090d56)

BibTeX

@misc{dc40bb5d439f04,
  title = {Can Benign Skills Induce Misaligned Behavior in LLM Agents? — arXiv:2611.04299},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.04299},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_40bb5d439f04; sha256 706ddfae39d4935af7f555957f92c432fa51d4b1d0f972ebc83fa85962090d56},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.04299. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

LLM agents increasingly rely on reusable skills to plan tasks, invoke tools, recover from errors, and complete workflows without modifying model parameters. Although these skills are typically developed to enhance benign tasks, they may also make an agent more capable of executing procedures for harmful purposes. We study this risk, which we call skill-induced behavioral misalignment: adding benign skills may cause an agent to execute external actions that violate safety requirements. We first conduct a systematic study to validate the existence of this phenomenon. It compares agents with and without benign skills across five LLM backbones and three benchmarks under matched requests. On average, skill augmentation improves benign task accuracy by 10.3 percentage points (pp), but also increases unsafe action rate (UAR) by 10.3 pp. The regression also occurs with skills curated by experts. A paired trajectory analysis shows that 24 of 50 newly unsafe trajectories fail to complete the harmful objective, revealing violations that task completion alone would miss. We further disclose multiple factors and scenarios that could exacerbate this risk. (1) Evolving skills with benign feedback can amplify unsafe execution; (2) This regression is more pronounced when retrieved skills are matched to task and contain procedural guidance; (3) Distilling benign experience into reusable skills yield higher UAR compared with their source trajectories. (4) A guard that checks proposed tool actions before execution reduces violations while retaining gains on benign tasks. All these findings confirm the significance of skill-induced behavioral misalignment, and the urgent need for new mitigation solutions.

Tema degli esploratori: fiducia digitale e crittografia (elenco arXiv cs.CR, nuovi lavori del giorno).
Categorie arXiv: Cryptography and Security (cs.CR).
Pagina del lavoro: https://arxiv.org/abs/2611.04299 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-04T21:38:19.524Z dalla pagina https://arxiv.org/list/cs.CR/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).