#692 · Primary category: AI Agents & Automation

CORAL

agent-framework agent-orchestration agentic-ai ai-agents alpha-evolve autonomous-agents autoresearch claude-code code-generation codex coding-agent evolutionary-algorithm large-language-models llm-agents multi-agent open-source opencode scientific-discovery self-evolving self-evolving-agents

Open-source autoresearch powered by autonomous coding agents. Run Claude Code, OpenCode, and Codex with grading, shared knowledge, and multi-agent evolution. Accepted at COLM 2026.

Project last updated:08/23/26

GitHub Stars

935

Forks

119

Contributors

20

License

Apache-2.0

Why we included this project

Anyone running long, multi-step coding tasks, the kind that take hours or days rather than a single prompt, will get real mileage out of CORAL. You point it at a codebase, hand it a grader, and it launches Claude Code, OpenCode, Codex, or similar agents inside isolated workspaces, keeping the evaluator separate from the work and letting agents see each other's shared state. Instead of a single code-generation pass, every attempt gets graded and added to shared knowledge that later runs build on, so the group explores, branches, and converges on a better solution over time. The paper behind the project has been accepted at COLM 2026.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category