Brainstorming for Tasks

Convert vague requirements into concrete design through structured questioning and exploration.

Announce at start: "I'm using the knowns.task.brainstorm skill to explore approaches."

Core principle: UNDERSTAND THE PROBLEM BEFORE PROPOSING SOLUTIONS.

The Process

Phase 1: Discovery

One question at a time. Don't overwhelm with multiple questions.

Prefer multiple-choice when possible:

Which approach do you prefer?
A) Quick solution with trade-offs
B) Comprehensive solution, more effort
C) Something else (describe)

Questions to clarify:

• What problem are we solving?
• Who are the users/stakeholders?
• What are the constraints?
• What does success look like?

Phase 2: Research Existing Patterns

json
1// Search docs for related patterns
2mcp__knowns__search_docs({ "query": "<topic>" })
3
4// Check how similar things were done
5mcp__knowns__search_tasks({ "query": "<keywords>" })

Learn from history - completed tasks often contain implementation insights.

Phase 3: Explore Approaches

Present 2-3 options with trade-offs:

markdown
1## Option A: [Name]
2- **Approach**: Brief description
3- **Pros**: What's good
4- **Cons**: What's challenging
5- **Effort**: Low/Medium/High
6
7## Option B: [Name]
8- **Approach**: Brief description
9- **Pros**: What's good
10- **Cons**: What's challenging
11- **Effort**: Low/Medium/High

Lead with your recommendation and explain why.

Phase 4: Validate and Document

After agreement:

• Summarize the chosen approach
• Identify potential risks
• Define acceptance criteria

If creating a new task:

json
1mcp__knowns__create_task({
2  "title": "<title>",
3  "description": "Based on brainstorm: <key decisions>",
4  "acceptanceCriteria": ["Criterion 1", "Criterion 2"]
5})

If updating existing task:

json
1mcp__knowns__update_task({
2  "taskId": "$ARGUMENTS",
3  "description": "Updated based on brainstorm..."
4})

When to Use This Skill

Good candidates:

• Vague requirements ("make it faster", "improve UX")
• Multiple valid approaches exist
• Significant effort involved
• New territory for the project

Skip for:

• Clear, well-defined tasks
• Bug fixes with obvious solutions
• Simple additions following existing patterns

Red Flags

You're doing it wrong if:

• Proposing solutions before understanding the problem
• Asking too many questions at once
• Not researching existing patterns first
• Skipping trade-off analysis

Remember

• One question at a time
• Research existing patterns first
• Present options with trade-offs
• Lead with your recommendation
• Document the decision