brief

分类: 数据与AI | 上传者: stefans71 | 下载: 0 | 版本: v1.0（最新）

Set the context for an upcoming review. Run when the user types /brief, describes something they want reviewed, or wants to change what the board is focused on. Also triggered automatically when the user asks the board to review something in plain language.

更新日志: Source: GitHub https://github.com/stefans71/FrontierBoard

目录结构

当前层级: .claude/skills/brief/

📄 SKILL.md 3.5 KB

SKILL.md

---
name: brief
description: Set the context for an upcoming review. Run when the user types /brief, describes something they want reviewed, or wants to change what the board is focused on. Also triggered automatically when the user asks the board to review something in plain language.
---

# Board of Governance — Brief

Prepare the board for a specific review. Either activate an existing context or write a new one, then populate each agent's inbox with a tailored brief for this question.

---

## Step 1: Understand What's Being Reviewed

If the user typed `/brief` without context, ask:

> What do you want the board to review? Describe it in plain language — a file, a decision, a proposal, a document, a question.

If this was triggered by a plain language request (e.g. "have the board look at this pricing model"), you already know. Confirm back what you understood before proceeding.

---

## Step 2: Detect the Domain

Based on what the user described, identify which domain this falls into — software, business, HR, finance, or something else.

Check `board/{agent}/contexts/` for each agent to see what context files already exist.

If a matching context exists for all agents, tell the user:

> I have a [domain] context ready for all agents. Want me to use that, or tailor the brief more specifically to this particular question?

If no matching context exists, tell the user:

> I don't have a [domain] context yet. I'll write one now — it'll be saved for future reviews of this type.

---

## Step 3: Write or Activate Context

**If using an existing context:** Read it to understand the domain lens each agent will apply.

**If writing a new context:** For each agent, write a context file at `board/{agent}/contexts/{domain}.md`.

A context file gives the agent a domain-specific lens — what to look for in this type of question, what questions to ask, what a good finding looks like here. It does not change the agent's identity. It focuses their attention.

A software context for a skeptic asks different questions than a software context for a systems thinker. Write each one for that specific agent's angle applied to this domain.

---

## Step 4: Write the Review Brief

Write a review brief for this specific question. This goes in each agent's inbox.

The brief should cover:
- What is being reviewed — the actual content, file, document, or question
- The specific question the user wants answered
- Any context the agents need to know (background, constraints, what's already been decided)
- What a useful finding looks like for this question
- Any specific concerns the user has flagged

If the user provided a file or document, include its contents or path in the brief.

The brief is the same for all agents — independent review of the same question is the point. Each agent's identity and context shapes how they approach it differently.

---

## Step 5: Populate Inboxes

Write the review brief to each agent's inbox. Name it `REVIEW-{date}-{short-description}.md` — something recognisable.

Tell the user:

> Brief is ready in all inboxes. Want me to run the board now, or do you want to review the brief first?

If they want to run now, follow the `/run` skill. If they want to review first, show them the brief and wait.

---

## Step 6: Save the Context (If New)

If a new context was written, confirm it's saved in each agent's contexts folder. Tell the user:

> I've saved a [domain] context for your board. Next time you ask for a [domain] review, I'll use it automatically — or you can ask me to tailor it further.

登录后下载/点赞/收藏 ❤ 2 | ★ 0

brief

目录结构

SKILL.md

举报内容

提示