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.
CORAL is infrastructure for autonomous coding agents that run experiments on a codebase, have their work graded, share knowledge, and evolve solutions across repeated attempts. It is used for autoresearch workflows with agents such as Claude Code, OpenCode, Codex, Cursor Agent, and Kiro. The catalogue entries provide skills, agents, instructions, a plugin, and a hook for operating these workflows.
Latest release v0.7.21 · 6 Sept 2026
6 files for Claude Code, Codex and OpenCode: coral-debug, coral-extend, coral-new-task, promoting-dev-to-main and 2 more — 6,180 tokens loaded in every session.
AGENTS.md A 1,602 tok CLAUDE.md A 4,578 tok .claude/skills/coral-debug/SKILL.md A 129 tok .claude/skills/coral-extend/SKILL.md A 152 tok .claude/skills/coral-new-task/SKILL.md A 138 tok .claude/skills/promoting-dev-to-main/SKILL.md A 33 tok These files are Human-Agent-Society/CORAL's own configuration — they tell Claude Code, Codex and OpenCode how to work on this repository, so they are not mods to install elsewhere. Copy one as a starting point and replace the parts that are about this project.