Metraipolis › archive › dc_83d0ef76ed14

AuraForge: Scaling Security Supervision for Training Coding Agents — arXiv:2610.00850

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.00850
License (as archived)
CC0-1.0
Archived
2026-10-06T13:57:05.694Z
SHA-256 of the archived text
f344bde4c2c3620b20cc6b56ecd999b309d97be40396095fa9c80a2d56e92d03
Stable URL of this page
https://metraipolis.com/documento/dc_83d0ef76ed14

How to cite

AuraForge: Scaling Security Supervision for Training Coding Agents — arXiv:2610.00850 — arXiv (en), https://arxiv.org/abs/2610.00850 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_83d0ef76ed14 (Metraipolis archive dc_83d0ef76ed14, sha256 f344bde4c2c3620b20cc6b56ecd999b309d97be40396095fa9c80a2d56e92d03)
BibTeX
@misc{dc83d0ef76ed14,
  title = {AuraForge: Scaling Security Supervision for Training Coding Agents — arXiv:2610.00850},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.00850},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_83d0ef76ed14; sha256 f344bde4c2c3620b20cc6b56ecd999b309d97be40396095fa9c80a2d56e92d03},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.00850. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Coding agents are now proficient enough to generate complex software applications from a single prompt. As their capabilities have grown, human oversight has increasingly shifted from line-by-line code review toward hands-off evaluation of outcomes. However, recent studies have shown that such a transition exposes a critical risk: functional correctness alone does not guarantee a secure implementation. Despite growing attention to code security, training safer coding agents remains challenging because reliable security supervision is difficult to obtain at scale from real-world repositories. We introduce AuraForge to synthesize and validate executable security tests for training secure coding agents. Our approach combines attack-oriented test synthesis, language-extensible task construction, and safeguards against reward hacking. Using AuraForge, we construct AuraGym, a multi-language and multi-CWE executable training gym: 679 executable feature-implementation tasks from 344 real-world repositories across Python, JavaScript, and TypeScript, covering 177 CWE categories. On the subset with human-written security tests, AuraForge produces about 3 times as many test cases on average and reduces the false-positive rate by 83.23%, allowing alternative secure implementations to receive correct supervision. Training Qwen3.5-4B with synthesized security tests gains larger improvements than human-written security tests (average 19.7 FuncPass and 6.2 SecPass vs. 14.9 FuncPass and 4.4 SecPass) on three languages. These results demonstrate that AuraForge provides more diverse and reliable security supervision to train secure coding agents.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-02 in cs.CL).
Categorie arXiv: Cryptography and Security (cs.CR) ; Computation and Language (cs.CL); Computers and Society (cs.CY).
Pagina del lavoro: https://arxiv.org/abs/2610.00850 — 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 2026-10-06T12:15:34.814Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-10-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).