clarify-roadmap

Clarify Roadmap

This skill encodes a deep conversational clarification loop to run before roadmap research and generation, when the user provides a high-level project vision that needs structured decomposition. It is instructional -- the calling command drives the loop using its own AskUserQuestion tool. This skill does not spawn subagents.

When to invoke

Apply this checklist to the raw input. If any item is missing or unclear, invoke the loop:

• Vision -- what does the finished product/system look like? What problem does it solve?
• Target users -- who are the primary users? What are their key pain points?
• Constraints -- budget, timeline, team size, regulatory requirements, platform requirements, existing commitments
• Success metrics -- how will the user measure success? (adoption, revenue, performance, user satisfaction, etc.)
• Technical preferences -- preferred stack, hosting, existing infrastructure to integrate with
• Scope boundary -- what is explicitly out of scope for this roadmap?
• Phasing priorities -- what needs to ship first? Are there hard deadlines for any phase?
• Risk tolerance -- appetite for new/unproven tech vs. battle-tested? Tolerance for shipping incrementally vs. all-at-once?

If all eight are present, skip this skill and proceed directly to research. Running the loop on an already-clear input is a cost, not a benefit.

Opt-out signals the caller must honor:

• --skip-clarify anywhere in the arguments
• config.phases.roadmap_clarify === false in .soloflow/config.json or config/defaults.yaml

Anti-pattern

"The user already knows what they want."

A roadmap shapes months of work. The checklist is the judge, not your intuition. Cheap to run, expensive to skip.

Scope-decomposition gate (runs first)

If the input spans multiple independent products (e.g., "build a marketplace AND a separate analytics dashboard AND a mobile companion app"), stop immediately. Do NOT begin clarification on the whole thing.

Use AskUserQuestion:

• Question: "This vision covers multiple independent products. Which one do you want to roadmap first?"
• Header: "Pick one"
• Options: one per detected product, plus a free-form "something else" fallback.

Roadmap only the selected product. The others can be roadmapped separately later.

Note: Unlike clarify-idea, which decomposes subsystems, this gate decomposes at the product level. A single product with multiple subsystems is fine -- the roadmap generator will phase those into epics.

Clarification loop

Rules (hard):

1. One AskUserQuestion at a time. Never batch clarification questions. Each answer informs the next question.
2. Prefer multiple-choice. Offer 2-4 concrete, mutually distinct candidate answers, plus the built-in free-form fallback the tool provides. Avoid open-ended prompts unless the space is genuinely unbounded (e.g., "describe your vision").
3. Follow threads. An answer may reveal a new ambiguity -- follow it. Examples:
   • "We want real-time features" -> "What kind of real-time? Live collaboration, notifications, or data streaming?"
   • "We need it done fast" -> "What's the hard deadline? And what's the team size?"
   • "We're targeting enterprise" -> "What compliance requirements? SOC2, HIPAA, GDPR?"
4. Lightweight research is allowed, not required. One targeted WebSearch for unfamiliar domain terms or stack questions is fine to frame better candidate answers. Never use research as a substitute for asking the user.
5. Do not propose epics, phases, or architecture. This skill only captures the vision and constraints. Structuring is the roadmap-generator's job.
6. Probe for priorities. When the user lists multiple goals, always ask which matters most. A roadmap without priorities is a wishlist.

Keep going until the checklist at the top of this file is satisfied -- every item has a clear answer.

Readiness gate (HARD-GATE)

Once you believe the checklist is satisfied, present an explicit confirmation via AskUserQuestion:

• Question: "I have a clear picture of your vision, constraints, and priorities. Ready to generate the roadmap?"
• Header: "Ready?"
• Options:
  • "Generate roadmap" -- proceed
  • "Keep clarifying" -- loop returns to the clarification loop
  • "Cancel" -- abort; caller stops without writing anything

You MUST NOT proceed to research/generation until "Generate roadmap" is selected. This is a hard gate -- the anti-pattern is declaring readiness yourself and skipping the user confirmation.

If the user picks "Keep clarifying," return to the clarification loop. There is no retry limit; loop until the user confirms.

Output: roadmap brief

When the gate passes, produce a roadmap brief in this exact shape and hand it to the caller:

## Raw Input

{Verbatim $ARGUMENTS the user passed in}

## Clarification Transcript

- Q: {question 1}
  A: {user answer 1}
- Q: {question 2}
  A: {user answer 2}
{...}

## Synthesis

### Vision
{One paragraph restating what the finished product looks like and why it matters}

### Target Users
{Who they are, their pain points, what success looks like for them}

### Constraints
{Timeline, budget, team, platform, regulatory -- everything that bounds the solution space}

### Success Metrics
{How success will be measured, in order of priority}

### Technical Preferences
{Stack, hosting, integrations, non-negotiable tech choices}

### Scope Boundary
{What is explicitly out of scope}

### Phasing Priorities
{What ships first and why, hard deadlines if any}

### Risk Tolerance
{New tech appetite, incremental vs big-bang shipping preference}

The Synthesis sections are canonical. The raw input and transcript are context only.

Reuse

This skill is designed to be invoked by /soloflow:roadmap. It could also be reused by any future command that needs deep project-level questioning before planning.