You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Please consider adding these papers to the Research & Discovery section:
Agon: An Autonomous Large-Scale Omnidisciplinary Research System Built on Prompt Economy β https://arxiv.org/abs/2606.24177 β Research orchestrator built as a Claude Code plugin. Scientist/coder/auditor loops across 10+ disciplines, governed by Prompt Economy: 18 roles in 230.6 KiB of prompts. GitHub: https://github.com/AutoResearch-Factory/Agon
AutoNumerics: An Autonomous, PDE-Agnostic Multi-Agent Pipeline for Scientific Computing β https://arxiv.org/abs/2602.17607 β Multi-agent pipeline that autonomously designs, implements, and verifies numerical PDE solvers from natural language. GitHub: https://github.com/Daviddjddu/Autonumerics
I traced our Ouroboros submission to the August 23 crawl: it gave up on 2608.08311 after three classification failures and marked the paper as seen. Could you retry its classification when convenient? The relevant run is https://github.com/EuniAI/awesome-code-agents/actions/runs/32624889751. Thanks for maintaining the list.
Activity
Zhaoyang-Chu commentedon May 29, 2026
https://arxiv.org/abs/2604.24212
Zhaoyang-Chu commentedon May 31, 2026
https://arxiv.org/abs/2605.18747
Zhaoyang-Chu commentedon Jun 1, 2026
https://arxiv.org/abs/2605.30248
Zhaoyang-Chu commentedon Jun 6, 2026
https://arxiv.org/abs/2605.26563
Zhaoyang-Chu commentedon Jun 6, 2026
https://arxiv.org/abs/2605.26186
Zhaoyang-Chu commentedon Jun 6, 2026
https://arxiv.org/pdf/2411.10213
Zhaoyang-Chu commentedon Jun 6, 2026
https://arxiv.org/abs/2605.29442
Zhaoyang-Chu commentedon Jun 6, 2026
https://arxiv.org/abs/2604.00436
Zhaoyang-Chu commentedon Jun 6, 2026
https://arxiv.org/abs/2604.20779
Zhaoyang-Chu commentedon Jun 6, 2026
https://arxiv.org/abs/2603.24631
Zhaoyang-Chu commentedon Jun 6, 2026
https://arxiv.org/abs/2602.07900
Zhaoyang-Chu commentedon Jun 6, 2026
https://arxiv.org/abs/2605.26720
Zhaoyang-Chu commentedon Jun 6, 2026
https://arxiv.org/abs/2511.00197
Zhaoyang-Chu commentedon Jun 6, 2026
https://arxiv.org/abs/2601.17406
Zhaoyang-Chu commentedon Jun 6, 2026
https://arxiv.org/abs/2506.18824
116 remaining items
WhymustIhaveaname commentedon Jul 17, 2026
https://arxiv.org/abs/2504.16918
Zhaoyang-Chu commentedon Jul 19, 2026
https://arxiv.org/abs/2606.15932
Zhaoyang-Chu commentedon Jul 19, 2026
https://arxiv.org/abs/2606.24530
Zhaoyang-Chu commentedon Jul 19, 2026
https://arxiv.org/abs/2606.22883
Zhaoyang-Chu commentedon Jul 19, 2026
https://arxiv.org/abs/2607.08964
Zhaoyang-Chu commentedon Jul 29, 2026
https://arxiv.org/abs/2607.24653
Zhaoyang-Chu commentedon Jul 30, 2026
https://arxiv.org/abs/2607.16740
Zhaoyang-Chu commentedon Jul 30, 2026
https://arxiv.org/abs/2606.28480
WhymustIhaveaname commentedon Aug 2, 2026
Please consider adding these papers to the Research & Discovery section:
Agon: An Autonomous Large-Scale Omnidisciplinary Research System Built on Prompt Economy β https://arxiv.org/abs/2606.24177 β Research orchestrator built as a Claude Code plugin. Scientist/coder/auditor loops across 10+ disciplines, governed by Prompt Economy: 18 roles in 230.6 KiB of prompts. GitHub: https://github.com/AutoResearch-Factory/Agon
AutoNumerics: An Autonomous, PDE-Agnostic Multi-Agent Pipeline for Scientific Computing β https://arxiv.org/abs/2602.17607 β Multi-agent pipeline that autonomously designs, implements, and verifies numerical PDE solvers from natural language. GitHub: https://github.com/Daviddjddu/Autonumerics
razzant commentedon Aug 11, 2026
https://arxiv.org/abs/2608.08311
Zhaoyang-Chu commentedon Aug 13, 2026
https://arxiv.org/abs/2608.07645
Zhaoyang-Chu commentedon Aug 13, 2026
https://arxiv.org/abs/2608.08311
Zhaoyang-Chu commentedon Aug 13, 2026
https://arxiv.org/abs/2608.09802
burgerseater commentedon Sep 4, 2026
https://arxiv.org/abs/2608.28281
razzant commentedon Oct 28, 2026
I traced our Ouroboros submission to the August 23 crawl: it gave up on 2608.08311 after three classification failures and marked the paper as seen. Could you retry its classification when convenient? The relevant run is https://github.com/EuniAI/awesome-code-agents/actions/runs/32624889751. Thanks for maintaining the list.