菜单
Use when pulling the next Klausemeister inbox item into the processing slot and creating a branch. Matches the /klause-pull workflow.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Use when interactively defining a project or feature PRD from a rough idea, problem statement, or product opportunity. Guides phased multiple-choice discovery for problem framing, user journey, edge cases, creative alternatives, and final PRD drafting. Matches the /prd workflow.
Use when the user wants to schedule a Linear project's tickets across worktree queues, distribute work based on dependencies, or says "schedule this project" / "assign tickets to worktrees" / "plan the work" / "save this schedule". Takes a Linear project, builds a dependency-aware assignment plan, then lets the user save the plan, enqueue it immediately, or cancel.
Use when waiting for a Klausemeister ticket's PR checks, merging the PR, and moving the ticket to Done. Matches the /klause-babysit workflow.
Use when defining a Klausemeister Backlog ticket, assessing its definition depth, stamping a simple/medium/complex label, and moving it to Todo. Matches the /klause-define workflow.
Use when executing a Klausemeister Todo ticket, routing implementation by its simple/medium/complex label, and moving it to In Progress. Matches the /klause-execute workflow.
Use when advancing the current Klausemeister item by reading product state and dispatching to the next workflow command. Matches the /klause-next workflow.
| name | klause-pull |
| description | Use when pulling the next Klausemeister inbox item into the processing slot and creating a branch. Matches the /klause-pull workflow. |
Pull a Klausemeister inbox item into processing using the same workflow as the
/klause-pull command.
The canonical workflow lives in this plugin's command spec at
./command.md.
Read that file now, then follow it exactly. It is the source of truth for:
transition(command: "pull")If the command spec is unclear or conflicts with live MCP state, stop and ask the user rather than guessing.