Agent Skills: PM Discovery Orchestrator

PM Discovery Orchestrator

Routes product discovery tasks to specialized frameworks.

Framework Routing

| Task | Skill | When to Use | |------|-------|-------------| | Customer interviews | pm-customer-interviews | Conducting or synthesizing user interviews | | Assumption testing | pm-assumption-mapping | Validating product hypotheses | | Understanding needs | pm-jobs-to-be-done | Identifying why customers "hire" solutions | | Feature prioritization | pm-prioritization | RICE/ICE scoring, opportunity mapping | | Roadmap planning | pm-roadmap | Theme-based Now/Next/Later roadmaps | | Feature briefs & decisions | pm-one-pager | Lightweight briefs, decision records |

When to Use Each

Interview synthesis needed? → pm-customer-interviews
Testing assumptions? → pm-assumption-mapping
Understanding motivation? → pm-jobs-to-be-done
Prioritizing features? → pm-prioritization
Planning roadmap? → pm-roadmap
Early exploration or recording decisions? → pm-one-pager
Full feature specification? → /majestic:prd

Discovery Synthesis

After using individual frameworks, synthesize findings:

## Discovery Summary: [Feature/Initiative]

### What We Learned
1. [Key insight with evidence]
2. [Key insight with evidence]
3. [Key insight with evidence]

### User Segments & Their Jobs
| Segment | Primary Job | Pain Intensity | Size |
|---------|-------------|----------------|------|
| [Segment A] | [JTBD] | [1-5] | [N users] |

### Prioritized Opportunities
| Rank | Opportunity | Evidence | RICE |
|------|-------------|----------|------|
| 1 | [Opp] | [Quote/data] | [Score] |

### Recommended Next Step
**Do:** [Specific action]
**Test:** [What to validate]
**Success looks like:** [Measurable outcome]

### What We Still Don't Know
- [ ] [Open question to investigate]
- [ ] [Assumption still untested]

Anti-Patterns

| Anti-Pattern | Why It Fails | Instead Do | |--------------|--------------|------------| | Leading questions | Confirms bias | Ask open-ended | | Hypothetical pricing | People lie | Ask current spending | | Feature requests as truth | Solutions, not problems | Dig for need | | Small sample decisions | Anecdotes ≠ patterns | 5+ signals minimum | | Skip competitor research | Reinvent wheel | Research first |

Integration

• Before: Use problem-research for market pain points
• After: Use /majestic:prd for requirements documentation
• After: Use /majestic:plan for implementation planning