Expansion Evaluation Skill

Expansion Evaluation Skill

Purpose

This skill manages a structured, resumable process for evaluating SoNash expansion ideas from 21 modules (12 feature + 9 technical). It maintains state across sessions and enables flexible navigation between modules and ideas.

When to Use

• User wants to evaluate expansion ideas
• User says "expansion", "/​expansion", or references the expansion docs
• User wants to check progress on expansion evaluation
• User wants to resume a previous evaluation session

Command Reference

Decision Actions

Use with "/​expansion decide":

• accept [milestone] [reason] - Stage idea for ROADMAP (NOT pushed immediately)
• defer [reason] - Good idea, but not now
• reject [reason] - Doesn't fit vision/​scope
• merge [roadmap-item] [reason] - Already in ROADMAP, enhance existing
• discuss - Mark for further discussion

ROADMAP Integration Process

CRITICAL: Decisions are STAGED, not immediately pushed to ROADMAP.md.

Why Staged?

1. Prevents ROADMAP churn during evaluation
2. Allows batch review before committing
3. User maintains control over when changes are made
4. Enables course correction before integration

Workflow

1. During evaluation, accepted ideas are logged in tracker's "Staged for ROADMAP" section
2. User can review staged items anytime with "/​expansion status"
3. User explicitly requests "/​expansion push-to-roadmap" when ready
4. Only then are items added to ROADMAP.md and committed
5. After push, staged items move to "Accepted Ideas Summary"

Push Command

/​expansion push-to-roadmap

1. Read all items from "Staged for ROADMAP" section
2. Present summary for user confirmation
3. On confirmation, add items to appropriate ROADMAP.md sections
4. Move items from "Staged" to "Accepted Ideas Summary"
5. Commit both files with descriptive message

Foundational Decisions (Resolved 2026-01-21)

These decisions guide ALL module evaluations:

Architecture

• Offline priority: Per-feature decision (not blanket policy)
• Encryption: Mandatory maximum for all step work
• Native wrapper: Deferred (focus on PWA first)

Features

• Nashville scope: Nashville-first, abstracted (city as parameter)
• Sponsor model: Push only (sponsee sends snapshots)
• Meeting finder: Explore automation scripts

Technical Tooling

• IndexedDB: Dexie.js (rich queries, React hooks, encryption addon)
• PDF generation: @react-pdf/​renderer (React components, lazy-loaded)
• Analytics: Minimal custom (Tier 1 anonymous + Tier 2 opt-in)

Process

• Evaluation order: Hybrid dependency-grouped (7 phases)
• ROADMAP integration: Staged with explicit push

Approved Evaluation Order

7-phase dependency-grouped flow:

Core Workflow

Begin Session

/​expansion begin

1. Read docs/​EXPANSION_EVALUATION_TRACKER.md
2. Display current progress summary
3. Show Quick Resume context from previous session
4. Suggest next module based on evaluation order
5. Show any staged items awaiting ROADMAP push

Evaluate Module

/​expansion evaluate T4
/​expansion evaluate F1 5

1. Load the specified module content from expansion docs
2. Check ROADMAP overlap for each idea (FIRST!)
3. Present ideas one at a time with evaluation criteria
4. Discuss feasibility, dependencies, user benefit
5. NEW: If accepting/​deferring, discuss ROADMAP placement:
   • Which milestone? (M5, M6, new milestone?)
   • Which section/​feature group within milestone?
   • Relationship to existing items? (new/​enhances/​replaces)
   • Insert where? (after which item, or append to end)
6. Record decisions incrementally in tracker with full placement metadata
7. Stage accepted/​deferred items (do NOT push to ROADMAP yet)

End Session

/​expansion-evaluation end

1. Update tracker with all decisions made
2. Set "Quick Resume" section with context for next session
3. List any new open questions
4. Remind user of staged items count
5. Commit tracker changes to git

Evaluation Criteria

For each idea, assess:

Placement must be discussed for all accepted/​deferred items.

Presentation Format (REQUIRED)

CRITICAL: Use this exact format for every idea evaluation. Do NOT deviate.

### [ID]: [Idea Name]

**Description:** [One-sentence summary of what this is]

**The Feature:**

- [Bullet point explaining what it does]
- [Implementation details]
- [How it works]

**Cross-Reference (optional):** [Only include if applicable; otherwise omit this
entire section]

**Technical Implementation (optional):** [Only include if relevant; otherwise
omit this entire section]

**Trade-offs:**

- **Pro:** [Benefit 1]
- **Pro:** [Benefit 2]
- **Pro:** [Benefit 3]
- **Con:** [Challenge/​risk 1]
- **Con:** [Challenge/​risk 2]

**Options:**

1. Accept [Milestone] - [Brief reasoning]
2. Defer - [Brief reasoning]
3. Reject - [Brief reasoning]
4. Merge [with item] - [Brief reasoning]

**Placement Recommendation:** [If Accept/​Defer option is viable]

- **Milestone:** [Which milestone - M5, M6, new milestone?]
- **Feature Group:** [Which group within milestone - M5-F1, create new?]
- **Insert After:** [MILESTONE:MX, ITEM:TX.X, or END:MX]
- **Relationship:** [NEW, BUNDLED_WITH:<ID>, REQUIRES_NATIVE,
  FUTURE_ENHANCEMENT, MERGED_INTO:<ID>]
- **Rationale:** [Why this placement makes sense - dependencies, grouping logic]

**Recommendation:** [Accept/​Defer/​Reject] - [Your rationale explaining why this
is the best option]

**What's your decision?**

Presentation Template Rules

1. Always include: Description, The Feature, Trade-offs, Options, Recommendation, question. (Placement Recommendation is conditional, see rule 6)
2. Optional sections: Cross-Reference, Technical Implementation (use when relevant)
3. Trade-offs format: Start each with "Pro:" or "Con:" prefix
4. List at least 3 pros and 2 cons - Be thorough in analysis
5. Options: Present all 4 options (Accept/​Defer/​Reject/​Merge) with brief reasoning for each
6. Placement Recommendation: ALWAYS include for items that could be accepted/​deferred - show milestone, feature group, insert after, relationship, and rationale
7. Recommendation: CRITICAL - Always recommend a specific decision (Accept/​Defer/​Reject/​Merge) with clear rationale explaining why. This is YOUR expert recommendation to guide the user's choice.
8. End with question: "What's your decision?" to prompt user choice

State Management

Primary state file: docs/​EXPANSION_EVALUATION_TRACKER.md

The tracker maintains:

• Quick Resume section with last session context
• Foundational Decisions reference
• Approved Evaluation Order (7 phases)
• Module progress tables with phase assignments
• Staged for ROADMAP section with placement metadata:
  • Placement (milestone-feature, e.g., "M4.5-F1")
  • Insert After (which item or "Create new milestone")
  • Relationship (new/​enhances/​bundled with)
• Deferred Ideas with placement metadata (for eventual push)
• Decision log per module
• Cross-reference table (F↔T dependencies)

File Locations

Implementation Notes

1. Always read tracker first - Check current state before any action
2. Update tracker incrementally - Don't batch updates within a session
3. Stage, don't push - Accepted items go to staging, not ROADMAP
4. Discuss placement during evaluation - For every accept/​defer, discuss milestone, feature group, and insertion point BEFORE moving to next idea
5. Record full placement metadata - Placement, Insert After, Relationship columns must be filled for all accepted AND deferred items
6. Include deferred items in placement - They also go to ROADMAP (future milestones), not just accepted items
7. Preserve discussion context - Log key points in tracker
8. Track dependencies - Note when ideas depend on other modules
9. Allow flexibility - User can jump to any module at any time
10. Remind about staged items - On "/​expansion end", show count of staged items and placement summary