跳到正文
cccskills
免费在 GitHub 公开

analiz-task-spec

Use when opening an analiz task for the system-architect - the required fields, what the criteria may say, and what the description must hand over

中文简介正在准备中。当前显示原文说明。

查看安装方法

包含的文件(1)

  • SKILL.md4.0 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

Analiz Task Spec

create_board_task fields for analiz:

  • task_type: "analiz" — the exact tool argument is task_type, not type. Passing the wrong field name is silently dropped (the tool ignores unknown arguments) and the task is created as task_type: "task" instead, which breaks analiz filtering, routing and board semantics. Always pass task_type: "analiz" explicitly.

  • column: todo

  • assignee: system-architect (REQUIRED — pass assignee: "system-architect" to create_board_task). Without an assignee the architect is never dispatched and the task sits idle in todo. The architect is not subscribed to the todo column; it reaches an analiz task only via the assignee.

  • title: "Analiz: [what is being investigated]"

  • description: investigation goal, scope boundaries, available context (links, prior docs).

    • Relevant projects/repositories (required): name every project/repository you believe the work touches, so the system-architect knows what to clone and review. If you are unsure, say so — the architect will verify and pull anything you missed. Missing a repo here does not block the architect (it cross-checks the codebase indexes), but naming them speeds the analysis.
    • Design brief, if the work touches UI: carry the same block implementation-task-spec uses (audience/tone, content language, visual direction, references, states, primary action) so the architect's spec inherits real design direction instead of inventing it. Detect an existing design system first (get_design_system, see implementation-task-spec); if none exists, open it now with request_design_system (implementation-task-spec) and name its key in the description — it runs while the architect analyses, and the architect makes the UI tasks wait for it.
  • acceptance_criteria — statements about the CONTENT of the spec and the plan, which is the only thing this task delivers. Write what must be true of the documents, e.g.:

    • "The spec names every repository the change touches and, per repository, the files and interfaces that change."
    • "Each unit of work in the plan states its inputs, its outputs and how it is verified."
    • "The plan orders the units so nothing depends on work scheduled after it."
    • "Every option considered is recorded with the reason it was or was not chosen."
    • "Any question only a stakeholder can answer is listed with the decision it blocks."

    Board actions are NEVER criteria. Not "moved to analiz_review", not "attached with add_task_document", not "presented for approval", and above all not "implementation tasks are created" — those are the workflow around the card (the architect's routine, the human's approval, what happens after it), and none of them says whether the analysis is any good. Worse, they cannot be ticked before the hand-off that ends the task, so the criteria gate holds the card and it never reaches done. The architect knows the workflow from its own role; the criteria are where you say what the documents must contain.

    Do NOT ask for the documents to be committed to the repository: an analysis produces a decision, not code, and a docs/ commit means a branch and a pull request for a task that ships nothing.

This analiz task is the permanent home of the spec and the plan. Every implementation task opened from it carries derived_from: ["A-N"] pointing back here, which is how the developer working that task is handed these documents; list_task_documents on this key returns them to anyone at any time. Nothing is ever copied into a repository, so if this card's documents are wrong or missing, so is every task built from it.

After creation: tell stakeholder "Opened an analiz task for the system-architect to investigate [topic]. It will come back to you in Analiz Review to approve the plan before any implementation task is created."

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

Use when writing acceptance criteria for a task - express each as an observable Given/When/Then that QA can execute, including negative cases

中文简介正在准备中。当前显示原文说明。

makifbaysal/tasktrooper1182026年10月21日 更新

Use when the diff adds or changes an endpoint, resolver, RPC, job or query that takes an object id, a role check, a request binding or a tenant filter - BOLA/IDOR, function-level authorization, mass assignment and tenant scoping

中文简介正在准备中。当前显示原文说明。

makifbaysal/tasktrooper1182026年10月21日 更新

Use on every UI change - semantic HTML, labels for controls, keyboard-navigable dialogs/menus, visible focus, and never color as the only signal

中文简介正在准备中。当前显示原文说明。

makifbaysal/tasktrooper1182026年10月21日 更新

Use when a task changes any screen, form, dialog, menu or control - Lighthouse/axe scan of the changed screens, a keyboard walk, and the thresholds that fail a task

中文简介正在准备中。当前显示原文说明。

makifbaysal/tasktrooper1182026年10月21日 更新

How to work a task returned with review, QA or UAT findings. Use when a task is in need_revision or PR review comments are in your context.

中文简介正在准备中。当前显示原文说明。

makifbaysal/tasktrooper1182026年10月21日 更新

Use when the repository keeps its alerting in code, or a person asks for alerts outside TaskTrooper - PrometheusRule YAML, a Cloud Monitoring alert policy in Terraform, a Datadog monitor, each routed to TaskTrooper's alert ingest URL so it opens an incident here

中文简介正在准备中。当前显示原文说明。

makifbaysal/tasktrooper1182026年10月21日 更新

查看 makifbaysal 的全部技能

报告此技能的问题