Nw DivergeSAFE

AI agents that guide you from idea to working code, with you in control at every step.

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: nw-diverge
description: "Generates 3-5 divergent design directions through JTBD analysis, competitive research, structured brainstorming, and taste evaluation before convergence. Use when the team has a validated problem but hasn't chosen a solution approach."
user-invocable: true
argument-hint: '[feature-id] - Optional: --work-type=[new-product|brownfield|pivot]'
---

# NW-DIVERGE: Structured Divergent Thinking Before Convergence

**Wave**: DIVERGE (between DISCOVER and DISCUSS, optional) | **Agent**: Flux (nw-diverger) | **Command**: `/nw-diverge`

## Overview

Execute DIVERGE wave through Flux's 4-phase workflow: JTBD analysis|competitive research|structured brainstorming|taste-filtered evaluation. Transforms a validated problem into 3-5 concrete, taste-scored design directions so DISCUSS can converge on one with confidence.

DIVERGE is optional. Brownfield features with a clear direction may skip it (see skip checklist in design spec). New products and pivot decisions benefit most from structured divergence.

## Interactive Decision Points

### Decision 1: Work Type
**Question**: What type of work is this?
**Options**:
1. New product -- no prior solution exists, full divergence needed
2. Brownfield feature -- existing product, exploring approach alternatives
3. Pivot / redesign -- existing feature being reconsidered from scratch
4. Other -- user provides custom context

### Decision 2: Research Depth
**Question**: How deep should competitive research go?
**Options**:
1. Lightweight -- 3 competitors, known market
2. Comprehensive -- 5+ competitors including non-obvious alternatives
3. Deep-dive -- cross-category research, adjacent markets, academic references

## Prior Wave Consultation

Before beginning DIVERGE work, read SSOT and prior wave artifacts:

1. **SSOT** (if `docs/product/` exists):
   - `docs/product/jobs.yaml` -- validated jobs and opportunity scores
   - `docs/product/vision.md` -- product vision and strategic context
2. **Project context**: `docs/project-brief.md` | `docs/stakeholders.yaml` (if available)
3. **DISCOVER artifacts**: Read `docs/feature/{feature-id}/discover/` (if present)
   - `wave-decisions.md` -- validated assumptions and key decisions
   - `problem-validation.md` -- customer evidence grounding the problem

**Migration gate**: If `docs/product/` does not exist but `docs/feature/` has existing features, STOP. Guide the user to `docs/guides/migrating-to-ssot-model/README.md` and complete the migration first. If greenfield, DIVERGE will bootstrap `docs/product/jobs.yaml` with the validated job.

**READING ENFORCEMENT**: You MUST read every file listed in Prior Wave Consultation above using the Read tool before proceeding. After reading, output a confirmation checklist. Do NOT skip files that exist -- skipping causes options disconnected from evidence.

## Agent Invocation

@nw-diverger

Execute \*diverge for {feature-id}.

**Context Files:** see Prior Wave Consultation above + project context files.

**Configuration:**
- work_type: {Decision 1}
- research_depth: {Decision 2}
- output_directory: docs/feature/{feature-id}/

**SKILL_LOADING**: Before starting work, load your skill files using the Read tool from `~/.claude/skills/nw-{skill-name}/SKILL.md`. Skills encode your methodology -- without them you operate with generic knowledge only.

At the start of execution, create these tasks using TaskCreate and follow them in order:

1. **JTBD Analysis** — Load `jtbd-analysis` skill. Extract and elevate the job from the raw request or DISCOVER evidence. Produce job statements (functional + emotional + social) and ODI outcome statements. Gate: job at strategic or physical level (not tactical), minimum 3 ODI outcome statements produced.
2. **Competitive Research** — Invoke `nw-researcher` sub-agent for evidence-grounded competitive research. Map how existing products serve the validated job. Identify non-obvious alternatives. Gate: 3+ real competitors named, at least one non-obvious alternative, evidence quality confirmed.
3. **Brainstorming** — Load `brainstorming` skill. Frame HMW question, apply SCAMPER lenses, generate structurally diverse options. Gate: 6 options generated with diversity confirmed (mechanism, assumption, and cost structure differ across options).
4. **Taste Evaluation** — Load `taste-evaluation` skill. Apply DVF filter, score surviving options on 4 taste criteria with locked weights, produce weighted ranking and recommendation with dissenting case. Gate: all surviving options scored on all 4 criteria, recommendation traceable to scoring matrix, dissenting case documented.
5. **Peer Review** — Invoke `nw-diverger-reviewer` (Prism) to validate all 5 dimensions. Revise if needed (max 2 iterations). Gate: reviewer approval confirmed, handoff accepted by nw-product-owner.

## Success Criteria

- [ ] Job extracted at strategic or physical level (not tactical, not a feature description)
- [ ] Minimum 3 ODI outcome statements produced
- [ ] 3+ real competitors researched, at least one non-obvious alternative
- [ ] 6 structurally diverse options generated (different mechanism, assumption, cost)
- [ ] All surviving options scored on all 4 taste criteria with locked weights
- [ ] Recommendation traceable to scoring matrix (no "feels right" overrides)
- [ ] Dissenting case documented for second-place option
- [ ] Peer review approved by nw-diverger-reviewer
- [ ] Handoff accepted by nw-product-owner (DISCUSS wave)

## Next Wave

**Handoff To**: nw-product-owner (DISCUSS wave)
**Deliverables**: `recommendation.md` with explicit decision statement + supporting DIVERGE artifacts

## Wave Decisions Summary

Before completing DIVERGE, produce (or append to) `docs/feature/{feature-id}/wave-decisions.md`:

```markdown
# DIVERGE Decisions -- {feature-id}

## Key Decisions
- [D1] {decision}: {rationale} (see: {source-file})

## Job Summary
- Validated job: {job statement at strategic/physical level}
- ODI outcomes: {count} outcome statements

## Optio

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

What the source does

Filesystem

none-observed

Network

none-observed

Shell

none-observed

Dependencies

pinned

Secrets in source

none-found

Findings (0)

No findings outside the package's declared scope.

Gates applied: no_behavioural_pass.