其他技能
rho-task-map
Maintain a task map that shows work, evidence, decisions and the next step to a person.
YuLab-SMU · 已有静态检查记录
安装与使用
运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。
npx skills add YuLab-SMU/Rho --skill rho-task-map技能说明
name: rho-task-map description: Maintain a task map that shows work, evidence, decisions and the next step to a person.
Rho task map
Describe the whole task yourself: Rho fixes no closed list of kinds or states.
- task.map.declare carries stable item IDs, your own kind and state words (each a label, an optional tone and an optional role Rho understands: working, waiting, done, planned, unknown, stopped), titles you write, progress counted from real receipts, an optional layout view per item, and next only for steps you will take.
- task.map.present changes a card form, its parameters (a subset), or the focus. People use the same call. A pinned target, or a change within agentFormIntervalMs of your last applied form change, is merged into one pending suggestion instead of being applied: the screen stays steady and the person applies it.
- task.map.respond is the person only: pin, unpin, revert your form change (layout only, never the work), accept or dismiss a suggestion.
- task.map.read observes. Reading never starts, resumes or reruns work.
- Declare the map when you start, not when you are done, and declare again as the work moves: until then the person sees your calls but not their structure.
- Name what an item rests on. As soon as artifact.check verifies a result, give the item artifact {id, version}; use from for an item that branches from another. Rho links an item to objects only through these references, never through its title.
- Mark a line judgment: true when it is your reading of a result rather than something a record shows. It is shown as your judgment, never as verified.
A form declared zoom: false is reached only by naming it, so stepping up from a card lands on the summary, not on compare.