Metraipolis › archive › dc_55cbdd134817

MAGS: Multi-agent Auto-formalization Guarantees Safety for Agentic Outputs — arXiv:2609.19391

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.19391
License (as archived)
CC0-1.0
Archived
2026-10-10T03:19:31.711Z
SHA-256 of the archived text
d79145c637aa14e8ef93952639bdc91e3491871d9ad22d741d6c0b635e7ec303
Stable URL of this page
https://metraipolis.com/documento/dc_55cbdd134817

How to cite

MAGS: Multi-agent Auto-formalization Guarantees Safety for Agentic Outputs — arXiv:2609.19391 — arXiv (en), https://arxiv.org/abs/2609.19391 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_55cbdd134817 (Metraipolis archive dc_55cbdd134817, sha256 d79145c637aa14e8ef93952639bdc91e3491871d9ad22d741d6c0b635e7ec303)
BibTeX
@misc{dc55cbdd134817,
  title = {MAGS: Multi-agent Auto-formalization Guarantees Safety for Agentic Outputs — arXiv:2609.19391},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.19391},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_55cbdd134817; sha256 d79145c637aa14e8ef93952639bdc91e3491871d9ad22d741d6c0b635e7ec303},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.19391. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

LLM coding agents now generate complex programs at a scale that makes thorough human review increasingly difficult, raising the risk of safety and security failures. Common approaches, including fuzz testing, static analysis, and LLM-as-a-Verifier, can detect many failures but struggle to cover all possible edge cases. Formal verification addresses this by providing machine-checkable guarantees over specified properties, but traditionally demands substantial manual specification and proof engineering. We introduce a unified multi-agent framework, MAGS, that generates executable programs with formal safety guarantees, using Dafny as a verification-aware intermediate representation where safety properties can be mechanically checked. MAGS formalizes and freezes human-audited APIs and safety requirements, translates generated code into Dafny, repairs violations using verifier feedback, and compiles verified programs back into executable code. We evaluate MAGS on 100 CUDA kernels, 100 terminal scripts, and 20 robotic-arm tasks. Across all 220 examples, it achieves a 100% success rate in producing programs with non-trivial safety guarantees against frozen specifications. Independent safety and functional evaluations further show strong performance across all three domains, while revealing failures when the auto-formalized semantics do not fully capture the target behavior.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-18 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Cryptography and Security (cs.CR); Software Engineering (cs.SE).
Pagina del lavoro: https://arxiv.org/abs/2609.19391 — 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 2026-10-10T03:04:13.596Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-18?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).