Use this page when you want to share Dayan Agent Skills without overstating what the repository proves.
Dayan Agent Skills is a MIT-licensed control library for AI agent work: 56 installable Skills, 15 public methods, fixtures, verifiers, and explicit evidence boundaries in one repository.
Prompts say what to ask. Control layers say what must be true.
Repository: https://github.com/Kosmoray/dayan-agent-skills
Most AI agent projects recreate the same missing layer around the model: task framing, permission boundaries, evidence contracts, stop conditions, and verification.
Dayan Agent Skills packages that layer as one open repository. It includes 56 installable public beta Skills, 15 public method cards, 12 copyable fixtures, and local verifiers that state what they prove and what they do not prove.
Start with the visual positioning demo, then install one Skill in a temporary home. If the library is worth revisiting, star it. If something fails in a clean environment, open a compatibility report.
I open-sourced Dayan Agent Skills: 56 reusable AI workflow Skills in one MIT-licensed repo.
The repo is not a prompt collection. It is a control library: workflow contracts, fixtures, verifiers, evidence boundaries, and known non-claims.
I am looking for clean-environment compatibility reports from Codex and Claude Code users.
Repository: https://github.com/Kosmoray/dayan-agent-skills
I open-sourced Dayan Agent Skills: 56 Skills and 15 public methods for giving AI agent work a control layer.
Prompts say what to ask. Control layers say what must be true.
MIT repo: https://github.com/Kosmoray/dayan-agent-skills
我把 Dayan Agent Skills 开源成了一个母舰仓库：56 个可安装 Skill、15 张公开方法卡、12 个可复制 fixture，还有本地验证器和证据边界。
它不是提示词合集，而是 AI 工作流控制层：先把任务、权限、证据、停止线和验收说清楚，再让模型工作。
仓库：https://github.com/Kosmoray/dayan-agent-skills
Ask people for one of these:
