shipkit-ideator - Shipkit Opportunity Ideator

Purpose: Turn gap analysis into actionable improvement ideas — new skills, workflow enhancements, and composition patterns

What it does:

• Reads the analyst report (docs/development/analyst-report.json)
• Brainstorms opportunities across four dimensions: new skills, skill improvements, workflow compositions, and architecture patterns
• Scores each opportunity on impact vs effort
• Produces ranked opportunity cards at docs/development/opportunities.json

---

When to Invoke

User says:

• "What should we build next?"
• "Brainstorm improvements"
• "Run the ideator"
• "What opportunities exist?"
• "How can we improve Shipkit?"
• "Suggest new skills"

Automated trigger:

• After /shipkit-analyst produces a new report
• As the third step of a Shipkit self-improvement team

---

Prerequisites

Required:

• docs/development/analyst-report.json exists (run /shipkit-analyst first)

Helpful context:

• docs/development/scout-report.json — Raw CC findings for additional context
• Previous opportunities at docs/development/opportunities.json
• Design philosophy at docs/development/SHIPKIT-DESIGN-PHILOSOPHY.md
• Skill quality standards at docs/development/SKILL-QUALITY-AND-PATTERNS.md

---

Process

Step 1: Load Context

Read these files:

1. docs/development/analyst-report.json — Gap analysis (required)
2. docs/development/scout-report.json — Raw CC findings (for additional context)
3. docs/development/opportunities.json — Previous opportunities (to avoid duplicates)
4. CLAUDE.md — Framework development instructions (for the Skill Value Test)

Extract from analyst report:

• All info findings (feature opportunities)
• All warning findings (deprecation migrations)
• skillCoverage array (coverage gaps)
• summary.skillsCoverage.lowest (weakest skills)

Step 2: Apply the Skill Value Test

Before brainstorming, recall the fundamental rule from CLAUDE.md:

A skill is valuable if it:

1. Forces human decisions to be explicit
2. Creates persistence Claude lacks

A skill is redundant if Claude does it well without instruction.

Every opportunity must pass this test. If Claude can do it well without a skill, it's not a valid opportunity.

Step 3: Brainstorm — New Skills

For each feature opportunity from the analyst report, consider:

1. Does this CC feature enable a new skill that passes the Value Test?

   • Example: New memory field → skill for managing what agents remember across sessions (persistence Claude lacks)
   • Counter-example: New syntax highlighting → Claude already does this, no skill needed

2. Are there workflow gaps the analyst found?

   • Missing handoffs between existing skills
   • Manual steps that could be automated
   • Context that gets lost between sessions

3. Are there team composition patterns that need skills?

   • Team orchestration variations beyond /shipkit-team
   • Specialized review patterns
   • Debug investigation workflows

For each new skill idea, produce:

{
  "type": "new-skill",
  "name": "shipkit-{proposed-name}",
  "category": "Vision|Planning|Knowledge|Execution|Quality|System",
  "description": "What it does",
  "valueTestResult": "Forces explicit: X / Creates persistence: Y",
  "triggeredBy": "GAP-xxx from analyst report",
  "ccFeatures": ["Features it would use"],
  "impact": 1-5,
  "effort": 1-5,
  "sketch": "2-3 sentence implementation approach"
}

Step 4: Brainstorm — Skill Improvements

For each existing skill with low coverage score:

1. What CC features is it not using that it could benefit from?

   • context: fork for isolated execution
   • memory for cross-session persistence
   • hooks in frontmatter for lifecycle events
   • New tool capabilities

2. What analyst findings apply to this skill?

   • Deprecated patterns to migrate
   • New patterns to adopt
   • Missing integrations

For each improvement idea:

{
  "type": "skill-improvement",
  "skill": "shipkit-{name}",
  "currentCoverage": 0.6,
  "proposedCoverage": 0.85,
  "improvements": [
    {
      "change": "Add context:fork to frontmatter",
      "benefit": "Isolated execution prevents context pollution",
      "effort": 1
    }
  ],
  "triggeredBy": "GAP-xxx",
  "totalImpact": 3,
  "totalEffort": 2
}

Step 5: Brainstorm — Workflow Compositions

Look at the full Shipkit pipeline and consider:

1. Which skill sequences could be automated with teams?

   • spec → plan → implement is already covered by /shipkit-team
   • What about: scout → analyst → ideator → plan → implement?
   • What about: feedback-bug → spec → plan → implement → verify?

2. Which skills work better in parallel than serial?

   • Multiple reviewers with different lenses
   • Multiple researchers evaluating options
   • Multiple implementers on different file clusters

3. What team templates are missing?

   • Compare against install/skills/shipkit-team/references/TEAM-TEMPLATES.md
   • Are there common workflows not covered?

For each composition idea:

{
  "type": "workflow-composition",
  "name": "Descriptive name",
  "pipeline": ["skill-a", "skill-b", "skill-c"],
  "parallelSteps": [["skill-b", "skill-c"]],
  "teamTemplate": "Template description if applicable",
  "triggeredBy": "Workflow gap observation",
  "impact": 4,
  "effort": 3,
  "sketch": "How the composition works"
}

Step 6: Brainstorm — Architecture Patterns

Consider broader framework improvements:

1. Are there new CC primitives that change how Shipkit should be structured?

   • New agent capabilities → rethink agent roles?
   • New hook events → new quality gates?
   • New settings options → new configuration patterns?

2. Are there cross-cutting concerns?

   • All skills could benefit from pattern X
   • The hook system could be extended with Y
   • The .shipkit/ context model could evolve

For each architecture idea:

{
  "type": "architecture-pattern",
  "name": "Pattern name",
  "scope": "framework-wide|skills|agents|hooks|settings",
  "description": "What the pattern is",
  "affectedComponents": ["List of affected files/systems"],
  "triggeredBy": "CC evolution or gap observation",
  "impact": 5,
  "effort": 4,
  "sketch": "Implementation approach",
  "risks": "What could go wrong"
}

Step 7: Score and Rank

For all opportunities:

1. Impact (1-5): How much does this improve Shipkit?

   • 5 = Enables entirely new capability
   • 4 = Significantly improves existing workflow
   • 3 = Noticeable improvement for users
   • 2 = Minor quality-of-life improvement
   • 1 = Cosmetic or marginal

2. Effort (1-5): How hard is this to implement?

   • 1 = Single file change, < 1 hour
   • 2 = 2-3 files, < half day
   • 3 = 4-6 files, ~1 day
   • 4 = Major feature, multi-day
   • 5 = Architecture change, multi-week

3. Priority Score = Impact / Effort (higher = do first)

4. Rank all opportunities by priority score

Step 8: Write Opportunities Report

Write docs/development/opportunities.json:

{
  "$schema": "shipkit-artifact",
  "type": "opportunities-report",
  "version": "1.0",
  "generatedAt": "2026-02-20T...",
  "source": "shipkit-ideator",
  "basedOn": {
    "analystReport": "docs/development/analyst-report.json",
    "analyzedAt": "2026-02-20T...",
    "ccVersion": "2.1.34"
  },
  "summary": {
    "totalOpportunities": 15,
    "byType": {
      "new-skill": 4,
      "skill-improvement": 6,
      "workflow-composition": 3,
      "architecture-pattern": 2
    },
    "topPriority": {
      "name": "Highest priority opportunity",
      "score": 2.5,
      "type": "skill-improvement"
    },
    "quickWins": 5,
    "majorInvestments": 2
  },
  "opportunities": [
    {
      "id": "OPP-001",
      "type": "new-skill|skill-improvement|workflow-composition|architecture-pattern",
      "name": "...",
      "impact": 4,
      "effort": 2,
      "priorityScore": 2.0,
      "category": "quick-win|standard|major-investment",
      "... type-specific fields ...": "..."
    }
  ],
  "quickWins": ["OPP-001", "OPP-003"],
  "roadmap": {
    "immediate": ["OPP-001", "OPP-002"],
    "shortTerm": ["OPP-005", "OPP-008"],
    "longTerm": ["OPP-012", "OPP-015"]
  }
}

Step 9: Present Summary

After writing the report, present a human-readable summary:

## Shipkit Improvement Opportunities

Based on CC v{version} analysis:

### Quick Wins (high impact, low effort)
1. OPP-001: {name} — Impact: 4, Effort: 1 → Score: 4.0
2. OPP-003: {name} — Impact: 3, Effort: 1 → Score: 3.0

### Standard Improvements
3. OPP-005: {name} — Impact: 4, Effort: 3 → Score: 1.3

### Major Investments
4. OPP-012: {name} — Impact: 5, Effort: 4 → Score: 1.25

### Suggested Roadmap
- **Now**: Quick wins (OPP-001, OPP-003)
- **Next sprint**: Standard improvements (OPP-005, OPP-008)
- **Later**: Major investments (OPP-012)

Ready to plan implementation? Run `/shipkit-plan` with any opportunity.

---

Output Quality Checklist

Before writing the report, verify:

• Every opportunity passes the Skill Value Test
• No duplicate ideas (check against previous opportunities.json)
• Impact/effort scores are justified, not arbitrary
• Each opportunity has a concrete sketch (not vague)
• triggeredBy traces back to specific analyst finding or gap
• Quick wins are genuinely quick (effort 1-2)
• Major investments have risks documented
• Roadmap is realistic (not everything in "immediate")

---

When This Skill Integrates with Others

Before This Skill

• /shipkit-analyst — Produces the gap analysis this skill reads
  • Trigger: Analyst report must exist
  • Why: Ideation needs to be grounded in actual gaps, not speculation

After This Skill

• /shipkit-plan — Plan implementation of a chosen opportunity
  • Trigger: User selects an opportunity to implement
  • Why: Opportunities need concrete implementation plans
• /shipkit-spec — Spec out a new skill opportunity
  • Trigger: User selects a "new-skill" type opportunity
  • Why: New skills need proper specification before building

Team Composition

In a self-improvement team:

• Scout runs first (intelligence gathering)
• Analyst runs second (gap mapping)
• Ideator runs third (this skill — opportunity generation)
• Output feeds into normal Shipkit workflow: spec → plan → implement

---

Context Files This Skill Reads

• docs/development/analyst-report.json — Gap analysis (required)
• docs/development/scout-report.json — Raw CC findings (additional context)
• docs/development/opportunities.json — Previous opportunities (dedup)
• docs/development/SHIPKIT-DESIGN-PHILOSOPHY.md — Design principles
• docs/development/SKILL-QUALITY-AND-PATTERNS.md — Quality standards
• CLAUDE.md — Skill Value Test and framework rules
• install/skills/shipkit-team/references/TEAM-TEMPLATES.md — Existing team templates

Context Files This Skill Writes

• docs/development/opportunities.json — Ranked opportunity cards

---

Mode Variations

Mode	What it does
--focus new-skills	Only brainstorm new skill ideas
--focus improvements	Only improve existing skills
--focus compositions	Only workflow composition ideas
--focus all	All dimensions (default)
--top 5	Only output top 5 by priority score
--top 10	Output top 10 (default)