菜单
Converse with the user to identify the next most productive units of work for the current project, and record them in project.md under Pending Work. Use when the user wants to plan what to work on next, or when there is no project.md yet.
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
| name | brainstorm |
| description | Converse with the user to identify the next most productive units of work for the current project, and record them in project.md under Pending Work. Use when the user wants to plan what to work on next, or when there is no project.md yet. |
| allowed-tools | Read(*) Write(project.md) Edit(project.md) |
Conversational skill. Examines the project state and helps the user decide what to do next. The only artifact this skill writes is project.md (creating it if missing, otherwise updating its Pending Work section).
project.md (optional). If missing, this skill bootstraps it from the conversation.work/index.yaml (optional). Lists previously-recorded work items.work/<slug>/README.md (optional). The current state of completed and pending work.project.md with structure:
# Goal
<multi-paragraph statement of the overall research/engineering goal>
# Background
<context, constraints, prior work, data sources>
# Completed Work
- [<slug>](work/<slug>/README.md) — <one-line summary>
# Pending Work
- [<slug>](work/<slug>/README.md) (status: <pending-plan|ready|...>) — <one-line summary>
Read state. Open project.md if it exists, plus work/index.yaml and any READMEs it references. If project.md is missing, ask the user about the goal and background and draft a first version — do not invent goals or background.
Converse. Discuss what would be most productive to do next given the goal, what is already complete, and what is pending. Ask clarifying questions when scope is ambiguous. Surface tradeoffs rather than picking unilaterally.
Propose work items. Each new unit of work should be small enough that a single planning pass can describe it concretely. For each:
work/)Confirm before writing. Show the proposed Pending Work changes to the user and wait for explicit approval.
Write artifacts. For each approved item:
project.md:
- [<slug>](work/<slug>/README.md) (status: pending-plan) — <one-line summary>work/<slug>/README.md with frontmatter and a Goal section only — leave Instructions, Results, Assessment empty for later skills to fill in:
---
slug: <slug>
status: pending-plan
---
# Goal
<multi-paragraph goal description>
# Instructions
# Results
# Assessment
Show the user the agent's work on a research project and save iterations on the user's behalf. Scaffold rendering and deploy infrastructure (Quarto today, GitHub Pages, dev container), show the rendered output, save iterations. Doesn't handle research execution (use `asta-flows`).
Install or update the `asta` CLI to the version pinned by this plugin. Invoke whenever `asta` is not on PATH (e.g. `command not found` for `asta`) or when a subcommand/flag appears missing — that signals a stale install needing an update.
Execute the approved plan in work/<slug>/README.md (status ready), write outputs into work/<slug>/data/, record what happened in the Results section, then hand off to review-work. Use when there is a ready work item.
Write a detailed, executable plan for a single unit of work, then get it approved by review-plan. Updates work/<slug>/README.md with concrete instructions and sets status to ready (or needs-input if review repeatedly rejects).
Critic skill. Given a plan in work/<slug>/README.md (status pending-review), check that the steps are concrete, the tools are appropriate, and the plan is realistic. Write a verdict and feedback back into the README.
Critic skill. Given an executed work item (status pending-assessment), assess whether the Goal was actually accomplished. Identify the root cause when it was not (incorrect assumption, technical blocker, missing data). Updates work/<slug>/README.md Assessment section.