Skip to content
View kubbot's full-sized avatar
🤒
哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈
🤒
哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈

Organizations

@kubecub @openkf @openim-sigs @aicworld @SoraEase @awesome-sora

Block or report kubbot

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
kubbot/README.md
kubbot AI workflow: a clear intent and relevant context feed agent execution; connected tools support the work; delivery includes a change and its evidence.

把想法变成可以检查的成果。这里展示我组织 AI 开发的方式:用上下文约束任务,让 Agent 调用工具推进工作,再用测试、审查和实际验收判断结果。

执行流程 · 上下文架构 · 搭建入口

One task, end to end

每一步交出具体产物:验收标准、隔离分支、验证证据,最后是可审查的 PR。检查失败就回到构建;发布和真实使用验收各自保留状态。

Task lifecycle: scope the goal and acceptance criteria, build in an isolated branch, verify with tests and review, then deliver a PR with evidence. Failed checks return to build. Release and real-world acceptance are separate checks.

Context that compounds

文档和仓库提供事实,Skills 说明方法,MCP 连接工具。记忆只保留经过核对的经验,下一次任务按需取用;重要取舍和高影响操作由人裁决。

Context architecture: sources, reusable skills, and reviewed memory supply relevant context to an agent. MCP connects tools. Verified work produces evidence, and reviewed outcomes feed lessons back into memory.

Build your own

组件 怎么接起来
任务入口 在 issue 中写清目标、约束和验收标准,再让 Agent 读取相关源码与官方文档。
可复用方法 从 Agent Kit 整理 Skills 和项目指令,让不同 Agent 使用一致的方法。
工具与执行 通过 MCP 连接需要的工具,在独立分支或 worktree 中实施变更。
验证与交付 用 GitHub Actions 跑适用检查,把改动、证据和未验证项一起放进 PR。
经验回流 核对实际结果后记录教训,让后续任务少走一次弯路。

架构图为本仓库维护的 SVG,支持明暗主题和窄屏布局。查看图源与生成方式。

Pinned Loading

  1. openimsdk/open-im-server openimsdk/open-im-server Public

    IM Chat OpenClaw

    Go 16.8k 3k