| name | mvp-roadmap-orchestrator |
| description | Use when turning a problem statement into an MVP plan, prioritization stack (Value/Effort, RICE, KANO, MoSCoW), 2-week sprint roadmap, PRD pack, and governed delivery documentation. |
| cluster | ai-agents-meta |
| version | 1.0.0 |
| origin | craft-agent workspace |
MVP Roadmap Orchestrator
Convert a raw idea into an execution-ready MVP plan with explicit prioritization, risk gates, scope control, and versioned product documentation.
Use When
Activate when the user asks for:
- problem statement → MVP
- roadmap, sprint plan, PRD, or delivery planning
- prioritization with Value/Effort, RICE, KANO, or MoSCoW
- POC before roadmap commitment
- 2-week sprint stories, Gantt/timeline, or scope governance
- documentation/versioning for product delivery
Out of Scope (Default)
Do not treat these as MVP by default unless explicitly requested:
- enterprise/compliance edge cases
- advanced automation and “nice-to-have” integrations
- multi-region scale architecture
- non-critical redesign work
Input Contract (Ask if Missing)
Before planning, gather:
- Target user + top pain
- Business objective + timeline
- Team capacity (roles, availability)
- Constraints (budget, legal, platform, dependencies)
- Success metric baseline (if available)
If key inputs are missing, ask targeted questions first.
Core Principles
- Problem-first, feature-second.
- Evidence over opinion.
- POC before full commitment on risky items.
- Scope discipline over roadmap vanity.
- Versioned docs and explicit decision ownership.
Execution Workflow
0) Discovery + Alignment
- Restate the problem and desired outcome in one paragraph.
- Confirm audience split: dev team, exec team, business, ops.
- Define planning horizon (MVP only vs MVP + next 2 horizons).
1) Problem Statement → Outcome Definition
- Capture pain, affected segment, current workaround, and impact.
- Define north-star metric + 2–4 leading indicators.
- Define MVP success threshold (launch is “done when …”).
2) Opportunity Backlog Creation
- Generate initiatives from user pain, business goals, and constraints.
- Add effort assumptions, dependencies, and risk notes per item.
- Mark each assumption as validated / unvalidated.
3) Prioritization Framework Stack (in order)
A) Value vs Effort
- Plot each opportunity: Quick Wins / Strategic Bets / Fill-ins / Avoid.
- Remove low-value high-effort items from MVP candidate list.
B) RICE
- Score formula:
(Reach × Impact × Confidence) / Effort.
- Show scoring ranges and assumptions for each variable.
- Flag low-confidence high-score items for validation.
C) KANO
- Classify as Must-have / Performance / Delighter.
- Keep Delighters out of MVP unless they unlock adoption risk.
D) MoSCoW
- Gate final sprint scope into Must/Should/Could/Won’t (this cycle).
- Explicitly list exclusions and rationale.
4) POC Gate (Before Roadmap Lock)
- Select top 1–3 risks (technical, desirability, viability, operational).
- Define each POC with:
- hypothesis
- success/fail criteria
- timebox
- owner
- decision date
- Require Go / No-Go / Pivot decision before roadmap commitment.
5) Research + Testing Loop
- Plan lightweight validation: interviews, surveys, prototype tests, concierge.
- Capture insight → decision mapping:
- Update RICE confidence and MoSCoW gates after each evidence round.
6) SLC Framing + Data Events
Define MVP through SLC:
- Simple: minimum viable user journey
- Lovable: moments that create confidence and repeat use
- Complete: end-to-end task completion for one core use case
For each SLC element, define data events:
- event name
- trigger point
- properties
- owner
- dashboard metric linkage
7) PRD Pack
Create PRD sections:
- problem/opportunity statement
- target users + JTBD
- scope in/out
- assumptions + dependencies
- functional requirements
- non-functional requirements
- acceptance criteria
- data/analytics plan
- risk register + mitigations
- launch readiness gates
8) Sprint Plan (2-Week Cadence)
- Build 2-week sprints with capacity-aware allocation.
- Write user stories using INVEST quality bar.
- Include acceptance tests for each story.
- Sequence work by dependency chain (product/design/dev/ops/legal if needed).
9) Timeline + Gantt View
- Convert sprint plan to timeline with milestones and decision gates.
- Mark critical path, parallel tracks, and buffer windows.
- Include rollback/contingency milestones for risky deliveries.
10) Scope Governance + Execution
- Use change requests for any new item after sprint lock.
- Re-score new work with RICE + MoSCoW before acceptance.
- Enforce WIP limits and prevent stealth scope creep.
11) Iterate + Version Documentation
- Run sprint review + retro with user and stakeholder feedback.
- Update roadmap as a time-based vision expansion.
- Version PRD, roadmap, events taxonomy, decisions, and release notes.
Hardened Controls (20 Improvements)
- Explicit input contract before planning.
- Mandatory clarification when critical fields are missing.
- Scope boundary section (Out of Scope) to prevent bloat.
- Assumption tagging as validated/unvalidated.
- Standard RICE formula and transparent scoring.
- Confidence-risk flags for low-confidence high-score items.
- Ordered framework stack to avoid random prioritization.
- Explicit exclusion list with rationale.
- POC Go/No-Go/Pivot gate before roadmap lock.
- Timeboxed POC with owner and decision date.
- Evidence update loop that re-scores priorities.
- SLC model tied to measurable data events.
- Event schema fields (name/trigger/properties/owner/metric).
- PRD minimum required sections defined.
- INVEST quality bar for stories.
- Acceptance tests required per story.
- Capacity-aware 2-week sprint planning.
- Critical path + contingency in timeline/Gantt.
- Change request + re-prioritization for scope creep.
- Documentation versioning + decision log ownership.
Output Artifacts (Default)
- Problem-to-MVP summary (with success criteria)
- Prioritized opportunity backlog (with Value/Effort + RICE + KANO + MoSCoW)
- POC plan with decision gates
- PRD draft or structured PRD outline
- 2-week sprint plan + user stories + acceptance tests
- Gantt-style timeline with milestones and critical path
- Scope governance policy (change control + prioritization rules)
- Documentation/versioning checklist + decision log template
Reusable Templates (Use by Default)
PRDTemplate.md
DecisionLogTemplate.md
BacklogScoringTemplate.md
When generating deliverables, instantiate these templates first, then fill with project-specific details.
NotebookLM Asset Automation (Python Module)
Use wired scripts to generate source-backed research artifacts and prefill planning docs:
- Wrapper (recommended):
run_mvp_pipeline.py
- Generator:
generate_assets_notebooklm.py
- Prefill:
prefill_templates_from_assets.py
- Guide:
NotebookLMIntegration.md
Default flow:
- Ensure auth (
notebooklm login).
- Run generator with sources (URLs/files) and selected assets.
- Run prefill script to produce PRD/backlog/decision drafts.
- Review and finalize with team-specific constraints.
Default command pattern:
./run_mvp_pipeline.py \
--project-dir /path/to/project \
--owner "<Owner>" \
--asset report --asset mind-map --asset data-table
Manual two-step fallback:
./generate_assets_notebooklm.py \
--title "MVP Research - <project>" \
--source "https://example.com" \
--source ./problem-statement.md \
--asset report --asset mind-map --asset data-table \
--output-dir ./outputs/<project>
./prefill_templates_from_assets.py \
--assets-dir ./outputs/<project> \
--templates-dir . \
--output-dir ./outputs/<project>/prefilled \
--project "<Project Name>" \
--owner "<Owner>"
Quality Gates (Definition of Done)
A plan is complete only when:
- success metrics are measurable,
- assumptions are explicit and risk-ranked,
- POC gates are defined for top risks,
- priorities include transparent scoring and exclusions,
- stories are testable and capacity-bounded,
- data events map to outcomes,
- and document ownership/versioning is clear.
Failure Modes to Avoid
- jumping into roadmap before defining MVP success
- scoring without confidence assumptions
- mixing Must-have and Delighter work in same sprint without rationale
- accepting scope changes without re-scoring
- claiming completion without testable acceptance criteria