> ## Documentation Index > Fetch the complete documentation index at: https://docs.tryhoard.com/llms.txt > Use this file to discover all available pages before exploring further. # Ask your AI assistant why cards did not reprice > Use a connected AI assistant to triage store-wide TCGplayer pricing exceptions, separate sync-wide blockers from individual card locks, and inspect the current rule without changing anything. When many cards appear not to have repriced, do not start by investigating cards one at a time. First find out whether one store-wide condition is affecting them all. This guide is a read-only triage of **your own Hoard store**: it does not change a rule, start a sync, unlock a card, roll back a price, or verify physical stock. First [connect your assistant](/talk-to-your-store). If you only have one suspicious listing, use [Ask your AI assistant why a card has that price](/how-to/ai-assistant/explain-a-card-price-with-ai) instead. ## Step 1: Diagnose the whole store first Ask for the store-wide result before naming individual cards: > "Why did my cards not reprice? Diagnose my whole store first. Show any system-wide warnings before the blocker counts, identify the largest blocker bucket, and give a few affected-card samples. Keep this read-only; do not trigger a sync or change anything." The assistant uses Hoard's `pricing.diagnoseStore` read to group cards by their current repricing state rather than making a separate request for every card. Expect counts and representative samples for buckets such as locked prices, a never-go-down floor, missing market reference, disabled updates, and cards already on target. Ask it to read the rule list separately for expired or disabled rules; the store diagnosis does not currently give a dedicated expired-rule warning. Read any system-wide warning before the counts. These can make every individual explanation secondary: * **Read-only mode** means Hoard can calculate prices but will not push them to TCGplayer. * **Agent offline** means a corrected or computed price cannot reach TCGplayer until the desktop agent is available. * **Stale sync** means recent calculations or pushes may not have landed yet. * **Auto-push hold** means a safety review is holding a batch. Review or reject that held batch in the Hoard dashboard; the assistant should not tell you to push around it. ## Step 2: Check sync health as a separate read Use the store diagnosis to see the pattern, then confirm how fresh the connection is: > "Check my sync health. Tell me the verdict, last successful sync, whether a sync is running, and the next safe action. Do not trigger a sync." Sync health reports what Hoard last observed. It does not prove that TCGplayer accepted every later update, and checking it does not repair an offline agent or release a held batch. Follow [check your sync health](/how-to/ai-assistant/check-tcgplayer-sync-health) for the meaning of each verdict. ## Step 3: Drill into one representative card Once you know the largest bucket, choose a sample from it—not every card in the store—and ask for its detailed explanation: > "For this sample card, explain why it has not repriced. Show its listed and target prices, governing rule, price lock or other current blocker, and sync context. Then evaluate the card against my current rules and compare that result with the stored target. Read only; do not apply a change." The assistant can use Hoard's `pricing.explainCard` read and `pricing.testCard` current-rule test for that listing. This distinguishes a one-card condition from a shared system condition: * A **price lock**, disabled price updates, missing market reference, or a never-go-down floor can affect one card or a subset of cards. * A read-only account, offline desktop agent, stale sync, or auto-push hold can affect the whole store or an entire batch. * A card that is **on target** has no price change to make, even if other cards need attention. For a locked card, use the Hoard dashboard to inspect or change the lock. The assistant cannot create or clear arbitrary per-card overrides through chat. ## Step 4: Do not confuse the saved receipt with today's rule The explanation may include a saved, materialized pricing receipt: the inputs and calculation steps Hoard recorded when it made an earlier pricing decision. That receipt is historical evidence, not a recalculation using today's market or rules. Ask the assistant to say whether the saved receipt still reflects the current price before treating it as today's explanation. A later manual edit, rollback, or sync without a matching receipt can make it stale. The current-rule test answers a different question: what the active rules calculate now. If the saved target and current calculation differ, treat that as a clue to check rule changes, market movement, and sync freshness—not as proof that either value has already been sent to TCGplayer. ## Step 5: Take the next action in the right place * For a shared warning, resolve the stated dashboard or desktop-agent issue before changing individual rules. * For repeated rule or floor behavior, ask the assistant to prepare a [pricing-rule proposal for review](/how-to/ai-assistant/propose-a-pricing-rule-matrix). Reviewing a diagnosis does not change prices. * For a one-card lock or manual exception, use the Hoard dashboard. Chat cannot directly override the card, unlock it, or roll it back. * If the diagnosis shows stale or failing sync, investigate the agent issue before requesting more work. A sync request is a separate write action; see [trigger a TCGplayer sync](/how-to/ai-assistant/trigger-a-tcgplayer-sync). Neither store diagnosis nor sync health confirms what is physically in stock. Reconcile quantity and condition separately before making fulfillment or inventory decisions. ## Related reading * [Explain a particular card's price](/how-to/ai-assistant/explain-a-card-price-with-ai) * [Check your TCGplayer sync health](/how-to/ai-assistant/check-tcgplayer-sync-health) * [Prepare a pricing rule matrix for review](/how-to/ai-assistant/propose-a-pricing-rule-matrix) * [What your AI assistant can and cannot do](/agent-capabilities)