| repo | MaximeRobeyns/self_improving_coding_agent |
|---|---|
| url | https://github.com/MaximeRobeyns/self_improving_coding_agent |
| content_timestamp | 2026-05-26 |
| time_slice | 2026-05 |
| timestamp_source | web_observed_github_search_result_2026_05_26 |
| collected_at | 2026-05-26 04:41:30 +0800 |
| source | github |
GitHub - MaximeRobeyns/self_improving_coding_agent: SICA implementation for a coding agent framework that works on its own codebase.
Source: https://github.com/MaximeRobeyns/self_improving_coding_agent
This raw-style public GitHub capture was recorded by the hourly public metadata update. Shell GitHub API freshness is not claimed for this capture; the source was a web-observed GitHub result.
- Repository: MaximeRobeyns/self_improving_coding_agent
- URL: https://github.com/MaximeRobeyns/self_improving_coding_agent
- Stars: 324
- Forks: 56
- Watchers: 11
- License: MIT
- Primary language / stack signal: Python 96.3%, JavaScript 2.6%, Dockerfile/Makefile/CSS/HTML remainder
- Collection timestamp: 2026-05-26T04:41:30+08:00
- The public GitHub page describes the project as a coding agent framework that works on its own codebase.
- The citation block identifies it as SICA, "A Self-Improving Coding Agent", associated with the ICLR 2025 Workshop on Scaling Self-Improving Foundation Models.
- The repository is core evidence for self-referential coding-agent improvement because the improvement target is the agent's own codebase, not only an external task answer.
No benchmark was run, no source clone was modified, and no private or authenticated metadata was used. This file preserves public page evidence for downstream classification, model-card analysis, public reports, and the site index.