opportunity-brief

Opportunity Brief

Produce a structured opportunity assessment that ranks plugin opportunities by impact and feasibility, with clear reasoning for what to build first. Designed for decision-makers — concise, evidence-based, actionable.

Expert Discovery

This skill uses dynamic expert loading. On every run:

1. Primary role: Always load product-manager.md (leads prioritization)
2. Scan project experts: Glob studio/agents/*.md — load all custom experts the team has created
3. Scan built-in experts: Glob ${CLAUDE_SKILL_DIR}/../../agents/*.md — load shipped experts (skip any already loaded from project)
4. Match by relevance: From the loaded experts, select those relevant to the current domain context. Match by comparing the expert's ## Your Domain section and title against the user's input topic. Include 1-3 most relevant domain experts.
5. Skip template: Do not load _domain-expert-template.md — it's for creating new experts, not for consultation.

The primary role produces the initial artifact. Domain experts review and correct it in the Expert Review step.

Inputs

Accept one of:

• A domain or project description via $ARGUMENTS
• A workspace path — read from event-storm.md (hotspots), personas/ (user impact), journeys/ (pain severity), domain-canvas.md (build strategy), behavior-matrix.md (automation potential)

The more prior artifacts exist, the more evidence-based the assessment.

Workflow

1. Collect evidence — gather pain points, hotspots, and opportunities from artifacts
2. Define candidates — list plugin opportunities with scope
3. Score — rate impact and feasibility
4. Estimate effort — rough complexity and timeline indicators
5. Rank and recommend — prioritized list with rationale
6. Expert review — domain experts validate feasibility and impact scores
7. Validate — present to user
8. Write output — save opportunity brief

Step 1: Collect Evidence

Gather from available sources:

Source	What to extract
event-storm.md	Hotspots with severity ratings
personas/*.md	Who is affected, how severely
journeys/*.md	Pain points per stage, emotional lows
processes/*.md	Decision points, manual steps, bottlenecks
domain-canvas.md	Core vs supporting classification
behavior-matrix.md	Automation opportunities, data gaps

If none of these exist, interview the user to gather equivalent information.

Step 2: Define Candidates

Each opportunity candidate is a potential plugin or feature:

ID	Candidate	Scope	Addresses
OP-1	智能膳食计划	根据目标和偏好自动生成周计划	HS-1 早餐纠结, PP-准备早餐
OP-2	快速饮食记录	语音/文字快速记录一餐	HS-2 记录成本高
OP-3	AI 营养顾问	即时个性化营养建议	HS-3 咨询慢
OP-4	运动方案适配	年龄段运动推荐	HS-4 运动不知选什么

Keep candidates focused — each should be one plugin or one major feature, not a whole platform.

Step 3: Score

Rate each candidate on two axes (1-5):

Impact — How much value does this deliver?

• 5: Daily use, eliminates a major pain point, affects all users
• 4: Frequent use, significantly reduces friction
• 3: Regular use, noticeable improvement
• 2: Occasional use, minor convenience
• 1: Rare use, minimal impact

Feasibility — How achievable is this?

• 5: Prompt-only skill, no external dependencies, proven pattern
• 4: Script-assisted, standard data processing
• 3: Needs domain knowledge base or moderate scripting
• 2: Needs MCP servers or external APIs
• 1: Needs complex orchestration, uncertain feasibility

        Feasibility →
    5   4   3   2   1
  ┌───┬───┬───┬───┬───┐
5 │★★★│★★★│★★ │★★ │★  │  ↑
  ├───┼───┼───┼───┼───┤  Impact
4 │★★★│★★ │★★ │★  │★  │
  ├───┼───┼───┼───┼───┤
3 │★★ │★★ │★  │★  │   │
  ├───┼───┼───┼───┼───┤
2 │★★ │★  │★  │   │   │
  ├───┼───┼───┼───┼───┤
1 │★  │★  │   │   │   │
  └───┴───┴───┴───┴───┘

★★★ = Build first   ★★ = Build next   ★ = Consider   (blank) = Defer

Step 4: Estimate Effort

For each candidate, provide rough indicators:

Candidate	Skills needed	Complexity tier	Dependencies	Effort hint
智能膳食计划	2	Moderate	营养数据库	中
快速饮食记录	1	Simple	—	小
AI 营养顾问	1-2	Simple	累计数据	小
运动方案适配	2	Moderate	运动知识库	中

Effort hints: 小 (1-2 skills, simple tier), 中 (2-4 skills, moderate tier), 大 (4+ skills, script-heavy or MCP).

Do NOT give time estimates — they're unreliable at this stage.

Step 5: Rank and Recommend

Produce the final priority ranking:

Priority Ranking
════════════════

 #  Candidate          Impact  Feasibility  Score  Effort
 1  快速饮食记录         5       5            25     小
    → 数据基础，其他一切都依赖它先存在

 2  AI 营养顾问         5       4            20     小
    → 高频痛点，替代人工咨询，prompt-only 可实现

 3  智能膳食计划         5       3            15     中
    → 核心价值，但需要营养数据库支撑

 4  运动方案适配         3       3             9     中
    → 有价值但非核心，可以后做

Recommendation:
  先建 #1 和 #2 — 两个小型 skill，快速验证价值
  然后建 #3 — 核心差异化功能
  #4 作为后续 add-on

Include a dependency note if candidates depend on each other (e.g., "AI 营养顾问 needs meal-log data from 快速饮食记录").

Step 6: Expert Review

If domain experts were discovered in Expert Discovery, use the Agent tool to have each relevant expert review the opportunity assessment.

Give each expert subagent:

• Their agent definition (.md file content)
• The draft opportunity brief (candidates, scores, ranking)
• The instruction: "Review this opportunity assessment from your domain expertise. For each candidate, evaluate: is the impact score realistic given domain constraints? Is the feasibility score accurate — are there domain-specific barriers we missed? Are there opportunities we overlooked that are obvious to a domain expert? Should any candidate be higher or lower priority based on your knowledge?"

Incorporate corrections. Common improvements from domain experts:

• Feasibility scores adjusted down due to regulatory barriers or domain complexity
• Impact scores adjusted up for candidates that address safety-critical pain points
• New opportunities that only a specialist would recognize
• Effort estimates corrected based on domain knowledge requirements

If no relevant domain experts were found, skip this step.

Step 7: Validate

Present the ranking to the user:

• "Do you agree with the impact and feasibility scores?"
• "Are there business constraints that change the priority?" (budget, timeline, team capacity)
• "Any candidates I missed?"

The user may override scores based on factors not visible in the data (e.g., "investors want to see the meal planning feature first").

Step 8: Write Output

If working within a studio workspace:

studio/changes/{domain}/opportunity-brief.md

If standalone, write to the current directory.

The file contains:

• Evidence summary (sources used)
• Candidate list with scope
• Impact × Feasibility scoring matrix
• Effort estimates
• Priority ranking with rationale
• Recommendation paragraph
• Dependencies between candidates