Brainstorming

Indexed Reference Map

• @[SKILL.md#L1-L94] Frontmatter, trigger rules, SOP, output contract.
• @[references/INDEX.md#L1-L17] Reading order and reference inventory.
• @[references/problem-framing.md#L1-L10] How to define the problem before proposing solutions.
• @[references/divergence-rules.md#L1-L7] How to explore multiple options without drifting.
• @[references/convergence-check.md#L1-L7] How to narrow to a recommended direction.
• @[references/examples.md#L1-L11] Example brainstorming outputs.

Scope

• In-scope:
  • Clarifying goals, constraints, unknowns, and trade-offs.
  • Generating multiple plausible approaches before planning.
  • Producing a recommendation with explicit reasoning.
• Out-of-scope:
  • Jumping straight into implementation details without framing.
  • Treating brainstorming as an excuse for endless ideation.
  • Replacing concrete planning when scope is already known.

KB Mode Notes

• Reading order:
  1. references/INDEX.md
  2. references/problem-framing.md
  3. references/divergence-rules.md
  4. references/convergence-check.md
  5. references/examples.md
• Coverage Matrix:
  • Problem definition -> problem-framing.md
  • Option generation -> divergence-rules.md
  • Narrowing decision -> convergence-check.md
  • Examples -> examples.md
• Portability note:
  • “Brainstorming” here is structured technical exploration, not free-form creative dumping.

Trigger Rules

• Positive:
  • User asks for ideas, options, approaches, design directions, comparisons, or trade-offs.
  • Problem framing is still fuzzy and solution space is open.
  • Prompt contains intent like: brainstorm, explore options, think through approaches, compare alternatives, what are the possible designs, before we implement.
• Negative:
  • A clear plan already exists.
  • The task only needs direct execution.
  • The user wants final verification/completion, not exploration.
• Escalation:
  • If brainstorming reveals the scope is already clear, hand off to writing-plans.
  • If unknowns are blocking good options, call them out instead of pretending certainty.

SOP

1. Frame the problem
   • Define objective, constraints, unknowns, and success criteria.
2. Generate candidate approaches
   • Produce multiple realistic options.
3. Compare trade-offs
   • Complexity, risk, reversibility, and fit with runtime constraints.
4. Identify missing information
   • Separate known facts from assumptions.
5. Converge to recommendation
   • Select the most sensible next direction.
6. Suggest next action
   • Usually move to plan-writing or targeted validation.

Error Handling

• Options are too vague:
  • restate the problem and tighten constraints.
• Brainstorming drifts into implementation:
  • pull back to approach-level comparison.
• Not enough context for good comparison:
  • say what is missing.
• No meaningful alternative exists:
  • state that directly instead of fabricating variety.

Output Contract

• Deliverable:
  • Problem Frame: objective/constraints/unknowns.
  • Options: 2+ candidate approaches when appropriate.
  • Trade-offs: pros/cons or risks.
  • Recommendation: preferred direction.
  • Next Step: usually plan, experiment, or direct execution.
• Validation:
  • Must distinguish assumptions from known constraints.
• Compliance Verdict:
  • PASS when exploration is bounded, comparative, and converges to a usable next step; otherwise PARTIAL or FAIL. herwise PARTIAL or FAIL.