Weekly Dealer Review

Dealer Group Support

If user_type is dealer_group:

1. Read dealer_group.locations[]
2. Ask: "Run weekly review for which location? Or 'all' for group rollup?"
   • Specific location: use that location's context
   • 'All': run per-location, then append GROUP WEEKLY ROLLUP

Group Weekly Rollup

IMPORTANT — Per-Location Stats Must Be Queried Individually: Each location row in the health scorecard MUST use its own API call with source=<location_domain> and stats=price,miles,dom. Never copy group-level stats across all rows — every location has different pricing, mileage, and DOM profiles. Query each location separately:

search_active_cars(source="location1.com", car_type="used", rows=1, stats="price,miles,dom")
search_active_cars(source="location2.com", car_type="used", rows=1, stats="price,miles,dom")
... one call per location domain

Then use each response's stats object for that location's row only.

GROUP WEEKLY ROLLUP — [Group Name]
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Location         | Units | Overpriced | At Market | Underpriced | Hot List Match | Stocking Gaps
-----------------|-------|-----------|-----------|-------------|---------------|-------------
[Location 1]     | XXX   | XX        | XX        | XX          | X of 10       | X categories
[Location 2]     | XXX   | XX        | XX        | XX          | X of 10       | X categories
...

GROUP STOCKING PRIORITIES:
  Most needed across group: [model 1], [model 2], [model 3]
  Best location for each: [which store's market has highest D/S ratio]

CROSS-LOCATION TRANSFER OPPORTUNITIES:
  [If one location is over-stocked in a category where another is under-stocked, flag it]

Execution: Multi-Agent Orchestration

Wave 1 — Launch Simultaneously

Launch these two agents in parallel using the Agent tool. Both are independent and can run at the same time.

Agent A: lot-scanner

Use the Agent tool to spawn the marketcheck-cowork-plugin:lot-scanner agent with this prompt:

Pull the complete inventory for dealer_id=[dealer_id], country=[US/UK], car_type=used, sort_by=dom, sort_order=desc. Paginate through ALL results — do not stop at 50. Return every vehicle with VIN, year, make, model, trim, listed price, mileage, and DOM.

Agent B: market-demand-agent (US only — skip for UK)

Use the Agent tool to spawn the marketcheck-cowork-plugin:market-demand-agent agent with this prompt:

Generate the stocking hot list and market demand snapshot for state=[state], dealer_type=[dealer_type], zip=[zip], radius=[radius], target_margin_pct=[target_margin], recon_cost=[recon_cost]. Use date range [first day of last month] to [last day of last month]. Run sections: hot_list, demand_snapshot.

Validate Wave 1

• lot-scanner returned pagination_status=complete — if partial, warn user in output
• Every vehicle row has vin, year, make, model, listed_price, dom
• market-demand-agent returned hot_list and demand_snapshot sections (US only) — if either missing, note in output

Wave 2 — After Lot Scanner Completes

Once the lot-scanner agent returns with the complete vehicle list:

Agent C: lot-pricer (US only)

Use the Agent tool to spawn the marketcheck-cowork-plugin:lot-pricer agent with this prompt:

Price these vehicles against the market: [pass the full vehicle list from lot-scanner]. Use zip=[zip], dealer_type=[dealer_type], floor_plan_per_day=[floor_plan_per_day], aging_threshold=[aging_threshold]. Price ALL vehicles — do not cap at 25.

UK Alternative (inline, no agent):

For UK dealers, instead of lot-pricer, price each unit inline:

• For each vehicle from lot-scanner, call mcp__marketcheck__search_uk_active_cars with matching year/make/model within radius, rows=10
• Calculate median price from comparables
• Classify as Below/At/Above Market using the same ±5% thresholds

Validate Wave 2

• lot-pricer returned pricing for all passed VINs — if some failed, note count in output
• No predicted_price = $0 or null in pricing output
• Every VIN in pricing output matches a VIN from lot-scanner inventory

Assembly — Combine Results

After all agents complete, assemble the report from their outputs:

1. Section 1 (Full Lot Competitive Scan) — from lot-pricer output:

   • Use the pricing table directly (already sorted by most overpriced first)
   • Use the summary statistics (above/at/below market counts and dollar estimates)
   • Each VIN appears once only (lot-scanner already deduped). If cross_listed_count > 1, show (×N sites) next to the source column as a note — do NOT repeat the row.

2. Section 2 (Stocking Hot List) — from market-demand-agent output (US only):

   • Take the hot_list top 10
   • Cross-reference with lot-scanner vehicle list: for each hot-list model, check if the dealer has any units. Flag gaps.

3. Section 3 (Market Demand Snapshot) — from market-demand-agent output (US only):

   • Use demand_snapshot top models and body type breakdown directly

4. Price Reductions — if querying price_change data for the group:

   • Deduplicate by VIN: same VIN reduced on multiple group sites = one entry, show the largest reduction amount. Note (×N sites) if cross-listed.
   • Never list the same VIN more than once in a price reduction table.

5. TOP 5 ACTIONS — synthesize from all agent outputs:

   • Priority 1-2: From lot-pricer (biggest overpriced units to reduce)
   • Priority 3: From lot-pricer (biggest underpriced units to raise)
   • Priority 4-5: From market-demand-agent (top models missing from lot to acquire)

Output Format

WEEKLY DEALER REVIEW — [Dealer Name] — Week of [Date]
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

FULL LOT COMPETITIVE SCAN — [N] Units Analyzed (all [total] units on lot)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

VIN (last 6) | Year Make Model | DOM | Your Price | Market Price | Gap % | Position | Action
-------------|-----------------|-----|------------|--------------|-------|----------|-------
[sorted by most overpriced first]

SUMMARY:
  [N] units ABOVE MARKET (avg [X]% overpriced) — reduce to recover ~$[X,XXX]
  [N] units AT MARKET — hold
  [N] units BELOW MARKET — consider raising [N] units to capture ~$[X,XXX]

STOCKING HOT LIST — Top 10 Models to Seek ([State], [Month])
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Rank | Make Model | Turn Days | Monthly Sold | Supply | D/S Ratio | Max Buy | On Your Lot?
-----|------------|-----------|-------------|--------|-----------|---------|-------------
[top 10 by opportunity score]

MARKET DEMAND — [State] — [Month Year]
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Top 10 Selling Models:
Rank | Make Model | Sold Count | Avg Price | Avg DOM
-----|------------|------------|-----------|--------
[table]

Demand by Segment:
Body Type | Sold Count | Share %
----------|------------|--------
[table]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
TOP 5 ACTIONS THIS WEEK:
1. [Highest-impact action with $ estimate]
2. [Second action]
3. [Third action]
4. [Fourth action]
5. [Fifth action]

Estimated total impact: $[X,XXX] in margin recovery + $[X,XXX] in stocking opportunity
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
For strategic monthly analysis (market share, depreciation, trends), run /monthly-strategy

UK dealers: Sections 2 and 3 are replaced with: "Hot List and Market Demand require US sold data. Use the competitive scan above for UK pricing intelligence."

Self-Check (before presenting to user)

• Tables have consistent column counts across all rows
• No $0 or null prices displayed
• Priced count matches scanned count (if gap > 5%, report pricing failures)
• D/S ratios calculated correctly (sold/supply, NOT supply/sold)
• Hot list max buy prices use the profile's margin% and recon cost
• Recommendations cite specific data points (not generic advice)
• Data period cited matches what was queried