# SpellBook Finance (remote · api.spellbook-finance.com) Magic: The Gathering card prices, movers, arbitrage, sealed box EV, and seller inventory tools. - Trust score: 66/100 (medium) - Registry status: active - Liveness: live - Owner verified: no - Last scored: 2026-09-28 ## Components - remote · `api.spellbook-finance.com`: 66/100 (this document), [markdown](https://verifymcp.io/servers/com-spellbook-finance-mcp/api.md), [page](https://verifymcp.io/servers/com-spellbook-finance-mcp/api) ## Channel facts - Endpoint: `https://api.spellbook-finance.com/mcp` - Transports: `streamable-http` - Auth: `none` - Version: `1.0.0` ## Trust breakdown How this component scores in each security and reliability category. Every signal is checked automatically against the live server, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog) Scored 2026-09-28. - **Endpoint Security**: 57/100 - The endpoint's TLS certificate is valid, in date, and uses a strong key. - Authorisation check failed: no authorisation is required to call this server, and it exposes a tool marked destructive (spellbook_ebay_publish). - HTTPS is enforced; there's no plaintext access path. - HSTS check failed: the Strict-Transport-Security header is absent. - DNSSEC check failed: this domain isn't protected by DNSSEC. - **Transport & Reachability**: 100/100 - Verified streamable-http transport via a live MCP handshake. - **Schema Quality & AI Usability**: 85/100 - 100% of prompts and resources have a non-trivial description (not blank, and not just the item's name). - AI-judged instruction clarity (excellent). - Context-footprint check failed: tool/resource definitions use about 21414 tokens (~131/item across 163 items; 100 tools + 63 resources), over budget; trim descriptions and params. - Usage-examples check failed: none of the tools include examples. - **Stability & Change Management**: 8/100 - Stability check failed: schema churn in the 3 days we've observed: 1 tool removals, 0 breaking changes, 0 auth/transport breaks, 36 additions. - **Tool Coverage**: 100/100 - 100% of tools have a non-trivial description (not blank, and not just the tool's name). - 100% of tool parameters carry a description. - **Tool Safety**: 100/100 - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured. - All 4 tool(s) whose name or description implies an irreversible operation declare an MCP destructiveHint annotation. - An AI judge read all 102 captured unit(s) of tool text and found none that tries to manipulate the model reading it. - **Capabilities**: 60/100 - Spec-recency check failed: implements MCP spec 2025-06-18; the latest is 2026-07-28. ## Install ### How do I install the SpellBook Finance MCP server? SpellBook Finance is a hosted endpoint at https://api.spellbook-finance.com/mcp, so there is nothing to install locally. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation. ### Claude ```bash claude mcp add --transport http com-spellbook-finance-mcp 'https://api.spellbook-finance.com/mcp' ``` ### Cursor ```json { "mcpServers": { "com-spellbook-finance-mcp": { "url": "https://api.spellbook-finance.com/mcp" } } } ``` ### VS Code ```json { "servers": { "com-spellbook-finance-mcp": { "type": "http", "url": "https://api.spellbook-finance.com/mcp" } } } ``` ### Codex ```toml [mcp_servers.com-spellbook-finance-mcp] url = "https://api.spellbook-finance.com/mcp" ``` ### opencode ```json { "$schema": "https://opencode.ai/config.json", "mcp": { "com-spellbook-finance-mcp": { "type": "remote", "url": "https://api.spellbook-finance.com/mcp", "enabled": true } } } ``` ### OpenClaw ```bash openclaw mcp add com-spellbook-finance-mcp --url 'https://api.spellbook-finance.com/mcp' --transport streamable-http ``` ### Hermes ```yaml mcp_servers: com-spellbook-finance-mcp: url: "https://api.spellbook-finance.com/mcp" ``` ### Netclaw ```json { "McpServers": { "com-spellbook-finance-mcp": { "Transport": "http", "Url": "https://api.spellbook-finance.com/mcp" } } } ``` ### Vellum ```bash assistant mcp add com-spellbook-finance-mcp -t streamable-http -u 'https://api.spellbook-finance.com/mcp' ``` ### Other ```json { "mcpServers": { "com-spellbook-finance-mcp": { "type": "http", "url": "https://api.spellbook-finance.com/mcp" } } } ``` The mcpServers block is a cross-client convention. Remote transports vary, so check your client's docs. ## Changelog Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims. ### 2026-09-28 (score 66, 0) - [functional] We updated how we score, so this day's move reflects our rubric, not a change to the server ### 2026-09-27 (score 66, +1) - [security regression] A breaking change shipped without a version bump: still 0.1.0 - [security] New tool “spellbook_sealed_portfolio_open”, which the server declares destructive - [security] Tool “spellbook_japan_source_scan” rewrote its description, which is the text the model reads - [security] Tool “spellbook_printings_search” rewrote its description, which is the text the model reads - [security] Tool “spellbook_sealed_listing_pause” rewrote its description, which is the text the model reads - [security] Tool “spellbook_sealed_listing_preview” rewrote its description, which is the text the model reads - [security] Tool “spellbook_sealed_listing_submit” rewrote its description, which is the text the model reads - [security] Tool “spellbook_sealed_portfolio” rewrote its description, which is the text the model reads - [functional regression] “spellbook_sealed_listing_submit” added a required parameter “expectedAskUsd”, so existing callers break - [functional] New resource “japanese-pokemon-sealed-box-shapes” - [functional] New resource “japanese-pokemon-single-language-prices” - [functional] New resource “mercari-japan-pokemon-landed-costs” - [functional] New tool “spellbook_sealed_recovered_component_record” - [cosmetic] “spellbook_pokemon_arbitrage_spreads” added an optional parameter “cohortId” - [cosmetic] “spellbook_pokemon_arbitrage_spreads” added an optional parameter “importChargesUsd” - [cosmetic] “spellbook_pokemon_arbitrage_spreads” added an optional parameter “importDutyRate” - [cosmetic] “spellbook_pokemon_arbitrage_spreads” added an optional parameter “inboundShippingUsd” - [cosmetic] “spellbook_pokemon_arbitrage_spreads” added an optional parameter “otherAcquisitionUsd” - [cosmetic] “spellbook_pokemon_arbitrage_spreads” added an optional parameter “otherSellingUsd” - [cosmetic] “spellbook_pokemon_arbitrage_spreads” added an optional parameter “outboundShippingUsd” - [cosmetic] “spellbook_pokemon_arbitrage_spreads” added an optional parameter “proxyFeeUsd” - [cosmetic] “spellbook_pokemon_arbitrage_spreads” added an optional parameter “sellingFeeRate” - [cosmetic] “spellbook_pokemon_arbitrage_spreads” added an optional parameter “sellingFixedFeeUsd” - [cosmetic] “spellbook_printings_search” added an optional parameter “language” - [cosmetic] “spellbook_sealed_listing_pause” added an optional parameter “channel” - [cosmetic] “spellbook_sealed_listing_preview” added an optional parameter “channel” - [cosmetic] “spellbook_sealed_listing_submit” added an optional parameter “channel” - [cosmetic] “spellbook_japan_source_scan” reworded the description of “cohortId” - [cosmetic] “spellbook_sealed_listing_pause” reworded the description of “idempotencyKey” - [cosmetic] “spellbook_sealed_listing_preview” reworded the description of “takeover” - [cosmetic] “spellbook_sealed_listing_submit” reworded the description of “idempotencyKey” - [cosmetic] “spellbook_sealed_listing_submit” reworded the description of “takeover” - [cosmetic] “spellbook_sealed_portfolio_hold_set” reworded the description of “entryId” ### 2026-09-26 (score 65, 0) - [security] New tool “spellbook_sealed_listing_pause”, which the server declares destructive - [security] New tool “spellbook_sealed_listing_submit”, which the server declares destructive - [security] Tool “spellbook_sealed_portfolio” rewrote its description, which is the text the model reads - [security] Tool “spellbook_sealed_product_search” rewrote its description, which is the text the model reads - [functional regression] Schema quality: 15971 → 18801 - [functional improvement] Stability: unverified → 0.03 - [functional] New tool “spellbook_card_buylist” - [functional] New tool “spellbook_deck_contents” - [functional] New tool “spellbook_deck_import_limitless” - [functional] New tool “spellbook_experiment_cross_channel” - [functional] New tool “spellbook_experiment_eligibility” - [functional] New tool “spellbook_pokemon_arbitrage_spreads” - [functional] New tool “spellbook_pokemon_ath” - [functional] New tool “spellbook_pokemon_market_analytics” - [functional] New tool “spellbook_pokemon_metagame_archetype” - [functional] New tool “spellbook_pokemon_metagame_card” - [functional] New tool “spellbook_pokemon_metagame_overview” - [functional] New tool “spellbook_sealed_intent_default_set” - [functional] New tool “spellbook_sealed_intent_defaults” - [functional] New tool “spellbook_sealed_listing_preview” - [functional] New tool “spellbook_sealed_portfolio_hold_set” - [functional] New tool “spellbook_sealed_portfolio_open_intent_clear” - [functional] New tool “spellbook_sealed_portfolio_receive” ### 2026-09-25 (score 65) First indexed and scored. ## MCP tools (100) ### `spellbook_guides` (~140 tokens) Find a cataloged SpellBook Finance customer-agent guide for seller setup and onboarding, selling, pricing, sealed EV, or choosing a card or sealed-product strategy for a budget and workload. Use it when a customer asks how to make money with Magic, what to do with a bankroll or store credit, whether to open or resell sealed product, or which research route fits their goal. Returns each guide's purpose, supported decision, and URL; read that page for its tool sequence and field meanings. Public search-acquisition articles are excluded. Input parameters: - `topic` (string): Narrow results to one education topic id. Omit to search every published topic. ### `spellbook_skill` (~150 tokens) Read the step-by-step procedure SpellBook Finance wrote for a whole customer task, in full. getting-started takes a new seller from sign-in through supplies, bin labels, a first tracked card, and connected marketplaces. buy-loop covers inspecting buy facts and recording what happened. physical-placement covers moving one item to a bin and verifying it. fulfilling-orders covers where an order lands, the bin walk that finds each sold card, and who packs, labels, and confirms shipped. Read the matching procedure before the first call of that task. Omit name to list every procedure with its title. Input parameters: - `name` (string): The procedure to read in full. Omit to list every procedure with its title. ### `spellbook_pull_work` (~216 tokens) Read a page of SpellBook's current fulfillment work, grouped by marketplace order. Each order carries its tray slot and card lines with bin, printing, condition, and state, including blocked reasons and ship-by evidence when available. Use when a seller asks what to pull or pack. Read spellbook_skill for fulfilling-orders before coaching the workflow; follow nextCursor until hasMore is false before summarizing the full returned workset. Use pending lines for the shared bin walk, then verify and pack each order separately. Input parameters: - `cursor` (string): Opaque nextCursor from the previous page. Pass it unchanged to continue; omit for the first page. - `limit` (integer): Maximum complete order groups to return on this page, from 1 to 50. Omit for 20. - `marketplace` (string): Filter to one marketplace, or use all for the combined bin walk. Omit for all. - `workspaceId` (string, required): The seller-owned workspace to read, from spellbook_workspaces_list. Required. ### `spellbook_buy_targets` (~298 tokens) Return a bounded source-order page of materialized TCGplayer-to-Card-Kingdom rows with an opaque nextCursor for complete traversal. Rows include printing identity, observed acquisition and cash quote prices, listed shipping, seller/quantity evidence, exit.quantityBuying (the copies Card Kingdom buys at the quoted price), maxCopies (the smaller of listed stock and exit.quantityBuying), the Card Kingdom condition ladder (NM/EX/VG/G payouts, modeled from the NM quote), timestamps, printingMismatches and missingFields. Call this first when a person says "I have $X to risk", "find me opportunities", "good opportunities", "single cards", "high risk singles", "start with arbitrage" or "what should I buy". Never request more copies of a row than maxCopies. Before handing a person a buy list, record each chosen row as a consideration with spellbook_customer_evidence_record, because spellbook_buy_outcome_report scores only recorded considerations. No profit, ranking, selection, reserves or cost defaults. Exact listing links and fee amounts are null when unavailable. Customer agent supplies all policy and executes purchases. Input parameters: - `cursor` (string): Opaque nextCursor from a previous page, passed back unchanged to continue traversal. Omit to start from the first page. - `limit` (integer): Maximum buy-target rows to return, 1 to 200. Omit for the default of 200. ### `spellbook_buy_outcome_report` (~89 tokens) Score every closed buy decision in one workspace against its immutable prediction and a fixed equal-weight printing basket. Includes unscored and abandoned counts, missing baseline evidence and small-sample limits. Uses submitted receipts; does not verify merchant orders or money. Input parameters: - `workspaceId` (string, required): The workspaceId to score, from spellbook_workspaces_list. Required; this scores exactly one workspace. ### `spellbook_printings_search` (~235 tokens) Find an exact card printing before recording a physical copy or forming a price. Searches by name, and narrows with `set:` plus a collector number in the same query string, for example "Pikachu set:sv2a 025/165". Pokemon searches can filter English or Japanese catalog rows. Returns each printing with its product identity, game, language, set code, collector number, rarity and available finishes. Read this first, because a card name alone does not identify a printing, and a wrong printing values the wrong card. Input parameters: - `game` (string): Restrict the search to one game. Omit to search Magic, the default. - `language` (string): Restrict Pokemon results to English or Japanese catalog rows. Omit when language is unknown. - `page` (integer): Page number of results, 1 to 20. Omit for the first page. - `q` (string, required): Card name to search, optionally narrowed with `set:` plus a collector number in the same string, e.g. "sol ring set:c21 263". Required. ### `spellbook_movers` (~177 tokens) Read which card printings changed price the most over one window: daily, weekly, 30d or 90d. Returns gainers and losers with exact game, catalog product, language, finish and marketplace variant when the source provides them, plus current and previous prices, dollar and percent change, a canonical SpellBook Finance cardUrl, and run timestamps. Call this to answer "what is moving this week" instead of guessing, then call spellbook_card_prices for one printing's per-source prices. An unavailable result names why no supported rows can be returned. Input parameters: - `game` (string): Restrict card movers to one game. Omit for the default Magic board. - `window` (string): Which price-change window to read: daily, weekly, 30d or 90d. Omit for the default window. ### `spellbook_box_ev_rankings` (~419 tokens) Read sealed-product EV rankings, defaulting to the highest modeled EV ratio first. Returns exact product IDs, market prices, modeled total EV, the acquirableFloorUsd and evRatioBasis used for the stored EV ratios when present, market-basis ratios, market-price spreads, shipping-aware ratios, per-pack metrics and row timestamps. evRatio is totalEV divided by max(marketPrice, the cheapest captured TCG listing price plus shipping) on rows with a known basis; evRatioMarketBasis preserves the market-price comparison. spreadUsd is totalEV minus marketPrice, so it is a separate comparison. This is research evidence for a caller's decision, not a guaranteed best box or a purchase instruction. Input parameters: - `box_type` (string): Exact box-type name to filter to one sealed product line. Omit to include every box type. - `cursor` (string): Opaque nextCursor from a previous page, passed back unchanged. Omit to start from the first page. - `game` (string): Restrict rankings to one game. Omit for the mixed supported-game ranking. A registered game without production EV returns an empty ranking plus sealedIntelligence.reason and missingFields. - `limit` (integer): Maximum rows to return, 1 to 150. Omit for the default of 50. - `max_price` (number): Maximum market price a row may have to be returned. Omit for no ceiling. - `min_ev_ratio` (number): Minimum stored evRatio a row must have to be returned. Rows with a null ratio are excluded. Omit for no floor. - `q` (string): Free-text filter on product name. Omit to include every product. - `sort` (string): Which computed field ranks the rows. Defaults to evRatio, the stored totalEV ratio using acquirableFloorUsd when its basis is known; use spreadUsd for totalEV minus marketPrice. - `sort_dir` (string): Sort direction for `sort`. Defaults to desc, highest first. ### `spellbook_card_prices` (~188 tokens) Read every price source SpellBook Finance holds for ONE printing, one row per source, each with its own label, price and timestamp. Sources can include TCGplayer Market, TCGplayer Low, Card Kingdom retail and buylist, CoolStuff, an eBay 90-day sold median, Cardmarket, ManaPool and CardTrader. Pokemon TCGplayer finish subtypes stay in separate rows. cardId is the exact printing id returned by spellbook_printings_search: a UUID for Magic, or a catalog product id for Riftbound or Pokemon. Rows come back in source order with their own timestamps; nothing is averaged, ranked, or chosen for you, and no listing link is returned. Input parameters: - `cardId` (string, required): The exact printing id from spellbook_printings_search: a UUID for Magic, or the namespaced catalog product id for Riftbound or Pokemon. Required. ### `spellbook_card_buylist` (~125 tokens) Read dealer buy offers for ONE exact card product, preserving each offer's game, finish, subtype, language, condition, source product, source timestamp, cash and credit fields, quantity and missingFields. Use the catalog product id from spellbook_printings_search. Pokemon offers are English NM/New CoolStuffInc observations when present; no offer or missing field proves a dealer has no stock. This read returns facts and does not choose a vendor or sell cards. Input parameters: - `cardId` (string, required): The exact printing or catalog product id from spellbook_printings_search. Required. ### `spellbook_pokemon_market_analytics` (~124 tokens) Read exact identity scoped Pokemon market facts for ONE catalog product: language, finish, condition, independent retail fair value, observed supply depth, versioned SCORE factors and Limitless demand coverage. The response keeps qualified values null when source count or comparable history is missing, and names missingFields and limitations. It reports facts only; it does not rank cards, recommend an action or infer a missing source. Input parameters: - `cardId` (string, required): The exact Pokemon catalog product id from spellbook_printings_search, such as pokemon:tcgp:12345. Required. ### `spellbook_pokemon_ath` (~106 tokens) Read the latest Pokemon all-time-high facts for the retained daily snapshot. Rows preserve exact product identity, language, finish, variant, source date, previous ATH and history coverage. This is a facts-only read; unavailable or partial history stays explicit in availability, missingFields and limitations. Input parameters: - `date` (string): Optional retained snapshot date, YYYY-MM-DD. Omit to read the latest snapshot. - `game` (string, required): The game to read. Must be pokemon. ### `spellbook_pokemon_arbitrage_spreads` (~344 tokens) Read bounded Pokemon acquisition and exit quote facts for exact product identity, preserving quote basis, currency, observation time, source reference, source URLs, comparisons, missingFields and limitations. This read compares observed sources; it does not rank opportunities, recommend a purchase, or execute a transaction. Input parameters: - `cohortId` (string): Optional Pokemon cohort. Use pokemon-ja or pokemon-en to keep language evidence isolated. - `condition` (string): Optional condition filter. - `finish` (string): Optional exact finish filter. - `importChargesUsd`: Optional import charges assumption in USD; omitted keeps landed cost incomplete. - `importDutyRate`: Optional import duty rate as a decimal; omitted keeps landed cost incomplete. - `inboundShippingUsd`: Optional inbound shipping assumption in USD; omitted keeps landed cost incomplete. - `language` (string): Optional exact language filter. - `limit` (integer): Maximum rows to return, from 1 to 100. Defaults to 25. - `otherAcquisitionUsd`: Optional other acquisition cost in USD; omitted keeps landed cost incomplete. - `otherSellingUsd`: Optional other selling cost in USD; omitted keeps net proceeds incomplete. - `outboundShippingUsd`: Optional outbound shipping assumption in USD; omitted keeps net proceeds incomplete. - `productId` (string): Optional exact Pokemon catalog product id from spellbook_printings_search. - `proxyFeeUsd`: Optional proxy fee assumption in USD; omitted keeps landed cost incomplete. - `sellingFeeRate`: Optional selling fee rate as a decimal; omitted keeps net proceeds incomplete. - `sellingFixedFeeUsd`: Optional selling fixed fee in USD; omitted keeps net proceeds incomplete. ### `spellbook_deck_contents` (~82 tokens) Read one authenticated customer's exact owned and draft card contents for a deck. The response preserves game identity, entry ids, matched catalog identity, quantities, membership recovery count and Pokemon readiness checks. It reports facts only and does not edit a deck or infer missing cards. Input parameters: - `deckId` (string, required): The exact deck id returned by the deck workflow. Required. ### `spellbook_deck_import_limitless` (~197 tokens) Import one player's Pokemon decklist from one fixed Limitless tournament standings record into the seller's workspace. The workspace owner chooses the event, player, format and idempotency key; retries with the same key replay the first result or report an in-progress import. The tool does not accept URLs or credentials and returns source identity, matched and unmatched counts, and the imported deck. Input parameters: - `eventId` (string, required): The Limitless tournament event id. Required. - `format` (string): Deck format. Defaults to Standard. - `idempotencyKey` (string, required): A caller-generated retry key. Reuse it unchanged after an uncertain result. Required. - `name` (string): Optional deck name. - `playerId` (string, required): The exact player name or id as shown by the selected event. Required. - `workspaceId` (string, required): The workspace to import into. The caller must be its owner. Required. ### `spellbook_pokemon_metagame_overview` (~97 tokens) Read observed Pokemon archetype coverage for one format from recent Limitless tournament decklists. The response preserves archetype facts, sample coverage, truncation, missingFields and limitations. It does not measure the whole player population or predict demand. Input parameters: - `format` (string, required): Pokemon format. Required. - `limit` (integer): Maximum archetypes to return, from 1 to 100. Defaults to 50. ### `spellbook_pokemon_metagame_archetype` (~119 tokens) Read observed Pokemon archetype and card facts for one format and exact archetype name from recent Limitless tournament decklists. The response preserves source coverage, card prices, truncation, missingFields and limitations. It reports observations and does not recommend a deck or purchase. Input parameters: - `archetypeName` (string, required): Exact observed archetype name. Required. - `format` (string, required): Pokemon format. Required. - `limit` (integer): Maximum cards to return, from 1 to 100. Defaults to 100. ### `spellbook_pokemon_metagame_card` (~120 tokens) Read observed Pokemon card demand facts for one exact catalog product, optionally narrowed to one format. The response preserves exact card identity, sample windows, counts, prices, coverage, missingFields and limitations. It reports observations and does not rank or recommend a card. Input parameters: - `cardId` (string, required): Exact Pokemon catalog product id from spellbook_printings_search. Required. - `format` (string): Optional Pokemon format filter. - `limit` (integer): Maximum observations to return, from 1 to 100. Defaults to 20. ### `spellbook_inventory_bin_ensure` (~230 tokens) Record physically confirmed cards into one inventory bin, by stating the DESIRED TOTAL quantity of each exact printing that belongs in that bin. Existing live copies count toward that total, so only a missing difference is created. Calling it twice with the same numbers creates nothing the second time. Use it for the first cards a seller enters by hand, with no CSV and no scanner. Returns the created item ids plus every candidate it refused and why. Input parameters: - `binId` (string, required): The bin these cards physically sit in. A bin id is a container letter plus an incrementing number, like B0001, where the letter names the physical box and the number names the bag inside it. Required. - `items` (array, required): Up to 200 exact printing rows to record in this bin. Required, at least one. - `physicallyConfirmed` (boolean, required): Must be true, asserting a person has physically counted these cards. The call is refused when this is omitted or false. - `workspaceId` (string, required): The workspace to record these cards into. Required; writes only to this workspace. ### `spellbook_inventory_import_csv` (~423 tokens) Import a seller's own ManaBox export or TCGplayer inventory export into one workspace, the same import the SpellBook web app runs. The server parses the CSV, resolves each row to an exact printing, assigns bins, and creates one inventory copy per card. TCGplayer Pokémon and Riftbound rows require the catalog's exact game, provider product identity, finish, condition, and language; incomplete or mismatched rows are skipped with a repair reason and never default to Magic, English, or nonfoil. Sending the identical CSV again returns the first result with duplicate true and creates nothing. Returns the importId, how many copies were written, the bins used, and every row it could not parse or skipped. Input parameters: - `binName` (string): A human name for that one bin, 1 to 48 letters, digits, spaces, dashes or underscores. Honored only when oneBinPerUpload is true. A name another import already holds is refused. - `binPrefix` (string): One capital letter that starts the auto-assigned bin codes, e.g. C gives C0001, C0002. Defaults to C. - `csv` (string, required): The full CSV text, header row included, exactly as the export produced it. Required. At most 200,000 characters. - `oneBinPerUpload` (boolean): True puts the whole file in one fresh bin instead of packing 200 cards per bin across the rolling sequence. Defaults to false. - `source` (string, required): Which export this is: manabox for a ManaBox CSV (needs Scryfall ID, Quantity, Condition and Foil columns), tcg for a TCGplayer seller inventory export (needs the TCGplayer Id column). Required. - `totalCost` (number): What the seller paid for the whole file in USD, split equally across every copy. Omit when the seller has not stated it; omitting records zero cost. - `workspaceId` (string, required): The workspace to import into. Required; writes only to this workspace. ### `spellbook_inventory_precon_import` (~360 tokens) Add one or more opened Commander precon decks to inventory as individual cards, the same Add Commander deck flow the SpellBook web app runs. The server reads the cached decklist for that TCGplayer product, creates one inventory copy per card at the exact deck printing, assigns bins, and splits the price paid across the cards. Returns the importId, how many copies were written, and the count placed in each bin. Input parameters: - `acquisitionDate` (string, required): The date the seller bought the deck, as YYYY-MM-DD. Required. - `binPrefix` (string, required): One capital letter that starts the auto-assigned bin codes, e.g. C gives C0001. Required. - `copyCount` (integer): How many identical decks were opened, 1 to 20. Defaults to 1. - `oneBinPerDeck` (boolean): True puts each deck in its own fresh bin. Defaults to false, which packs the cards into the rolling bin sequence. - `overrideDuplicate` (boolean): True imports the deck even when this seller already imported the same product. Set it only when the person confirms a second physical deck. - `paidPriceUsd` (number, required): What the seller paid for ONE deck in USD. Required. Send 0 only when the deck cost nothing. Every copy of the deck is recorded at this price. - `tcgProductId` (integer, required): Exact TCGplayer product id of the sealed precon deck, from spellbook_sealed_product_search. Required. A product that is not a single Commander or Riftbound precon deck is refused. - `workspaceId` (string, required): The workspace to add the cards to, from spellbook_workspaces_list. Required; writes only to this workspace. ### `spellbook_purchase_incoming_record` (~300 tokens) Record single cards the customer bought on a marketplace that have not arrived yet, one row per exact printing with the quantity and the price paid per copy. Each copy is written as pending receipt under one order, so it counts as incoming and cannot be listed or sold until it is received. Returns the order's acquisitionSource, the copy count, the total cost and the bins the copies will go into. Use it right after the customer completes a marketplace checkout for singles found through spellbook_buy_targets or spellbook_singles_buylist_spreads. It never buys anything and moves no money. A card already in hand goes through spellbook_inventory_bin_ensure instead. When the package arrives, call spellbook_inventory_incoming_receive with the returned acquisitionSource. Input parameters: - `marketplace` (string, required): Where the order was placed. Required. Use other for any marketplace not listed. - `orderRef` (string, required): The marketplace's order number from the checkout confirmation. Required. The same order number groups every copy for receiving. - `rows` (array, required): The order's lines, one per exact printing, condition and finish. Required, at least one. - `sellerName` (string): The seller's store name on the marketplace, when the checkout shows one. Shown beside the order in the incoming list. - `workspaceId` (string, required): The workspace that bought the cards, from spellbook_workspaces_list. Required; writes only to this workspace. ### `spellbook_inventory_incoming` (~133 tokens) Read what a workspace has bought or saved that has not arrived yet, one group per marketplace order or precon deck import. Each group carries its acquisitionSource, a readable title, the copy and card counts, the total cost, the date it was recorded and the bins its copies are assigned to. Call this to answer what is still on the way, or to find the acquisitionSource to receive. It reports pending groups; it does not track a package or say when one will arrive. Input parameters: - `workspaceId` (string, required): The workspace whose incoming inventory to read, from spellbook_workspaces_list. Required. ### `spellbook_inventory_incoming_receive` (~148 tokens) Mark every pending copy in the named orders or precon imports as received, which makes them owned inventory that can be listed and sold. Returns how many copies were received, their total cost and the bins they belong in. Call it when the customer says the package arrived and they counted the cards. Read spellbook_inventory_incoming first to get each acquisitionSource. Input parameters: - `acquisitionSources` (array, required): The acquisitionSource of each order or import to receive, exactly as spellbook_inventory_incoming or spellbook_purchase_incoming_record returned it. Required, at least one. - `workspaceId` (string, required): The workspace that holds the pending copies. Required; writes only to this workspace. ### `spellbook_ebay_publish` (~388 tokens) Publish owned copies from one workspace as live eBay listings, priced by a stated rule. Pass itemIds to publish one exact card. ALWAYS call it once with dryRun true and show the person the counts and the preflight before calling it again with dryRun false, because a live run creates real listings a real buyer can buy. Pricing anchors on market price, never on what the seller paid: multiplier 0.96 lists at 96% of market, and floorUsd only ever raises a price. A live run returns a queued jobId, or the reason nothing published. A queued job has listed nothing yet, so call spellbook_ebay_job_status with that jobId before telling the person any card is live. Input parameters: - `batchSize` (integer): Maximum copies to submit in this call. Omit for the default batch size. - `binId` (string): Restrict publishing to one exact bin. Omit to consider every owned copy. - `dryRun` (boolean): True previews the run with no live listings created. False publishes for real. Defaults to true; a live run needs an explicit false. - `floorUsd` (number): A price floor in USD that only ever raises a computed price. Omit for no floor. - `itemIds` (array): Restrict publishing to these exact Item ids, from a prior spellbook_inventory_items_query. Pass one id to list one card. The owned-only rule still applies to each id. Omit to consider every owned cop… - `multiplier` (number, required): Price multiplier applied to each copy's market price, e.g. 0.96 lists at 96% of market. Required. - `setCodes` (array): Restrict publishing to these set codes. Omit to consider every set. - `workspaceId` (string, required): The workspace whose owned copies to publish. Required. ### `spellbook_ebay_reprice` (~359 tokens) Preview or reprice existing live eBay listings in one seller workspace using a market multiplier and optional floor. This is a one-run marketplace override: it changes current eBay asks, creates no listing, and saves no pricing profile or durable override. Preview first with the default dryRun true; an explicit dryRun false queues an async reprice job. Read that job with spellbook_ebay_job_status until each affected listing reaches its terminal state. Input parameters: - `acquisitionSource` (string): Restrict the run to one exact acquisition source. Omit for every live eBay listing in the workspace. - `binId` (string): Restrict the run to one exact inventory bin. Omit for every live eBay listing in the workspace. - `dryRun` (boolean): True previews the listings and counts with no marketplace change. False queues the reprice job. Defaults to true; applying requires explicit false. - `floorUsd` (number): Optional minimum ask in USD. Existing backend floors and locks can still raise the resulting ask. - `itemIds` (array): Restrict the run to these exact listed Item ids from spellbook_inventory_items_query. Omit for every live eBay listing in the workspace. - `multiplier` (number, required): Market price multiplier for the selected listings, such as 0.96 for 96% of market. Required. - `setCodes` (array): Restrict the run to these set codes. Omit for every live eBay listing in the workspace. - `skipIfWithinPercent` (number): Skip an existing ask already within this percentage of the target. Defaults to 2. - `workspaceId` (string, required): The workspace whose existing eBay listings to reprice. Required. ### `spellbook_ebay_unpublish` (~141 tokens) End the live eBay listings for up to 25 exact copies, and keep each copy in inventory so it can be published again. Take Item ids from spellbook_inventory_items_query. Show the person which cards will come down and get a yes before calling, because a live listing ends at once. Returns one result per copy: unpublished, kept, no_listing, or failed. Input parameters: - `itemIds` (array, required): Up to 25 exact Item ids whose eBay listings to end, from a prior spellbook_inventory_items_query. Required, at least one. - `workspaceId` (string, required): The workspace that owns the copies. Required. ### `spellbook_ebay_job_status` (~152 tokens) Read one eBay bulk job by the jobId returned from spellbook_ebay_publish or spellbook_ebay_reprice. Both operations queue work, so call this before telling the person a listing was created or its ask changed. State planning, pending, uploading, processing or ingesting means the job is still running: call again after about a minute. State failed carries errorMessage, and each failures row carries a per-card reason: relay that reason to the person. Returns state, counts, per-card items, failures and missingFields. Input parameters: - `jobId` (string, required): The jobId returned by spellbook_ebay_publish or spellbook_ebay_reprice. Required. ### `spellbook_manapool_sync_preview` (~174 tokens) Compare one workspace's inventory against what is live on ManaPool, and return the exact changes a sync would make. Every action is listed with its kind: upsert to create or change a listing, delist to remove one, blocked where a cross-channel check refused it. Nothing is sent to ManaPool by this call. Show the person the summary and the counts, then call spellbook_manapool_sync_apply with the previewToken this returns. Input parameters: - `floorUsd` (number): A price floor in USD for this preview. Omit to use the workspace's stored floor. - `multiplier` (number): Price multiplier applied to market price for this preview. Omit to use the workspace's stored pricing. - `workspaceId` (string, required): The workspace to compare against live ManaPool state. Required. ### `spellbook_manapool_sync_apply` (~224 tokens) Send a previewed batch of changes to ManaPool, creating, repricing and removing real listings. Takes the previewToken from spellbook_manapool_sync_preview, which must still be valid. Applies at most maxActions in one call and returns remainingActionCount with nextActionCursor, so a full book takes several calls. Show the person the preview before the first call, not after. Input parameters: - `floorUsd` (number): Must match the floorUsd used to build the preview, or the token is refused. - `maxActions` (integer): Maximum actions to apply in this call, 1 to 250. Omit for the default of 50; call again with the same token to continue. - `multiplier` (number): Must match the multiplier used to build the preview, or the token is refused. - `previewToken` (string, required): The previewToken from spellbook_manapool_sync_preview. Must still be valid; a stale token is refused. - `workspaceId` (string, required): The workspace whose previewed changes to apply. Required. ### `spellbook_sealed_listing_preview` (~349 tokens) Preview one sealed listing on ManaPool or TCGplayer without publishing. Returns the proposed quantity and ask, any current offer, TCG publisher readback state, stock and intent revisions, and blockers. For TCGplayer, the first call may queue an exact seller-grid SKU read; call again after the paired extension finishes. After submitting, call preview again: publisher.state live confirms an offer, and removed confirms a zero-quantity removal. Pass this preview's channel, ask, and revisions to spellbook_sealed_listing_submit. Input parameters: - `boxId` (string, required): The TCGplayer product id of the sealed product, the boxId on a spellbook_sealed_portfolio row. Required. - `cap` (integer): The most units to offer at once. Omit to offer every eligible unit. - `channel` (string): Marketplace to preview. Defaults to manapool. - `fixedPriceUsd` (number): A fixed ask in USD for every unit. Omit to list at the product's current market price. - `game` (string): The game of the product. Defaults to mtg. - `kind` (string, required): The sealed kind, the kind on the same spellbook_sealed_portfolio row. Required, because units are counted by kind and boxId together. - `language` (string): The product language code, such as EN or JA. Defaults to EN. - `takeover` (boolean): True adopts a listing on the selected marketplace that SpellBook did not create. Defaults to false; pass true only after the person agrees. - `workspaceId` (string, required): The workspace that holds the sealed units, from spellbook_workspaces_list. Required. ### `spellbook_sealed_listing_submit` (~449 tokens) Submit one sealed product listing to ManaPool or TCGplayer at the exact quantity and ask shown by spellbook_sealed_listing_preview. Pass the same channel and product arguments, plus proposed.stockRevision, proposed.askUsd, and intentRevision unchanged. ManaPool submits directly. TCGplayer queues the paired browser writer. A queued TCG write is pending until an exact seller-grid readback confirms it. Input parameters: - `boxId` (string, required): The TCGplayer product id of the sealed product, the boxId on a spellbook_sealed_portfolio row. Required. - `cap` (integer): The most units to offer at once. Omit to offer every eligible unit. - `channel` (string): Marketplace to submit to. Must match the preview. Defaults to manapool. - `expectedAskUsd` (number, required): proposed.askUsd from the preview. Required; a different value means the market ask changed and the call is refused. - `expectedIntentRevision`: intentRevision from the preview. Pass it unchanged, null included. - `expectedStockRevision` (integer, required): proposed.stockRevision from the preview. Required; a different value means stock moved and the call is refused. - `fixedPriceUsd` (number): A fixed ask in USD for every unit. Omit to list at the product's current market price. - `game` (string): The game of the product. Defaults to mtg. - `idempotencyKey` (string, required): A unique request key for this publish attempt, such as a new UUID. If the result is uncertain, preview current state before deciding on another submit. Required. - `kind` (string, required): The sealed kind, the kind on the same spellbook_sealed_portfolio row. Required, because units are counted by kind and boxId together. - `language` (string): The product language code, such as EN or JA. Defaults to EN. - `takeover` (boolean): True adopts a listing on the selected marketplace that SpellBook did not create. Defaults to false; pass true only after the person agrees. - `workspaceId` (string, required): The workspace that holds the sealed units, from spellbook_workspaces_list. Required. ### `spellbook_sealed_listing_pause` (~238 tokens) Remove one SpellBook-managed sealed listing from ManaPool or set its TCGplayer quantity to zero through the paired browser. Call spellbook_sealed_listing_preview first and pass its channel and intentRevision unchanged. A product with no saved SpellBook intent cannot be removed by this tool. Input parameters: - `boxId` (string, required): The exact TCGplayer sealed product id from the portfolio row. Required. - `channel` (string): Marketplace whose managed listing should be removed. Defaults to manapool. - `expectedIntentRevision` (integer, required): The non-null intentRevision from a fresh spellbook_sealed_listing_preview. Required. - `game` (string): The product game. Defaults to mtg. - `idempotencyKey` (string, required): A unique request key for this removal, such as a new UUID. If the result is uncertain, preview current state before deciding on another pause. Required. - `kind` (string, required): The sealed kind from the same portfolio row. Required. - `language` (string): The product language code. Defaults to EN. - `workspaceId` (string, required): The workspace that owns the listing. Required. ### `spellbook_tcg_listing_preview` (~214 tokens) Show which cards in one workspace would be listed on TCGplayer, and at what price, without changing anything. TCGplayer has no server-side publish, so the real path is a CSV a person uploads to the TCGplayer portal. Call this first and show the person the candidates and the totals, then call spellbook_tcg_listing_export only once they have agreed to upload the file. Input parameters: - `cursor` (string): Opaque cursor from a previous page, passed back unchanged. Omit to start from the first page. - `floorPrice` (number): A price floor in USD for this preview. Omit to use the workspace's stored floor. - `marketMultiplier` (number): Price multiplier applied to market price for this preview. Omit to use the workspace's stored pricing. - `sampleSize` (integer): Maximum candidate rows to preview. Omit for the default sample size. - `workspaceId` (string, required): The workspace to preview a TCGplayer listing run for. Required. ### `spellbook_pokemon_tcg_staged_preview` (~133 tokens) Preview the TCGplayer CSV rows for 1 to 25 exact Pokemon physical inventory copies in one workspace. Call this after the seller selects the copies they want to inspect. It returns the existing staged-preview rows and CSV; it queues nothing, publishes nothing, and offers no publish or export follow-up. Input parameters: - `itemIds` (array, required): The exact Item ids selected by the seller, 1 to 25 unique non-empty values. Required; each copy must be owned and belong to Pokemon. - `workspaceId` (string, required): The workspace that owns every selected Pokemon inventory copy. Required. ### `spellbook_tcg_listing_export` (~189 tokens) Produce the CSV a person uploads to the TCGplayer portal, and mark those copies as listed in SpellBook. Call it only after spellbook_tcg_listing_preview and only once the person has said they will upload the file. Hand them the CSV and tell them plainly that SpellBook now believes these copies are listed, so the upload has to happen. Input parameters: - `floorPrice` (number): Must match the value shown in spellbook_tcg_listing_preview to export the same priced candidates. - `marketMultiplier` (number): Must match the value shown in spellbook_tcg_listing_preview to export the same priced candidates. - `sampleSize` (integer): Must match the value shown in spellbook_tcg_listing_preview to export the same candidate set. - `workspaceId` (string, required): The workspace to export a TCGplayer upload CSV for. Required; writes listing records for this workspace. ### `spellbook_manapool_connection_status` (~72 tokens) Check whether this signed-in customer's ManaPool seller account is connected to SpellBook Finance. Call this before requesting a setup link and again after the seller finishes the setup form. It returns a masked account identifier and stored connection timestamps. If connected is false, call spellbook_manapool_connect_link. ### `spellbook_manapool_connect_link` (~155 tokens) Get the SpellBook Finance page where the seller connects ManaPool, and guide them through it. Call spellbook_manapool_connection_status first. Save the seller's return address with spellbook_return_address_save before you hand over the link, because the page fills ManaPool's tax address from it. Hand the seller only this one link. The page itself links to ManaPool's token page, where a seller with no ManaPool account signs up first and then creates an access token starting mpat_. Tell them the SpellBook page already shows their email and return address, so they paste only the token. Call spellbook_manapool_connection_status after they finish. The agent never receives or submits the credential. ### `spellbook_item_photo_upload_link` (~183 tokens) Get a link where the seller uploads their own phone photo of one inventory copy, so the next eBay listing shows that photo in place of catalog art. Use it when the seller wants buyers to see the real card, or when catalog art is wrong or missing. Take itemId from spellbook_inventory_items_query or spellbook_inventory_holdings. Give the seller the returned url and tell them to open it on their phone, then take or pick a photo of the card front. No photo passes through you. After they say it is saved, call spellbook_ebay_publish for that copy. Input parameters: - `itemId` (string, required): One inventory copy's itemId, from spellbook_inventory_items_query or spellbook_inventory_holdings. Required. - `workspaceId` (string, required): The workspace that holds the copy. The caller must own it. Required. ### `spellbook_ebay_connection_status` (~98 tokens) Check whether this signed-in customer's eBay seller account is connected to SpellBook Finance. Call this before requesting a connect link and call it again after the seller approves eBay consent. It returns the stored account identity and token state plus current eBay seller-registration and selling-limit facts. If connected is false and account is null, call spellbook_ebay_connect_link. If status is revoked, fetch a fresh connect link. ### `spellbook_ebay_connect_link` (~108 tokens) Get a one-click link the seller opens to connect their eBay account to SpellBook Finance. Call spellbook_ebay_connection_status first. The link is signed for this customer and expires, so fetch a fresh one when you hand it over. Give them the link and tell them eBay will ask them to sign in and approve the connection. Once they finish, call spellbook_ebay_connection_status again; the agent never needs to open a SpellBook browser. ### `spellbook_tcg_extension_pairing` (~164 tokens) Read whether this workspace has a paired TCGplayer browser extension, and get the SpellBook page where the seller pairs one. TCGplayer has no seller API, so SpellBook syncs TCGplayer listings and orders through a Chrome extension on the seller's own computer. Returns whether the account has extension access, the current session's status, when it was paired, when the extension last checked in, and its order sync state. If session is null or lastSeenAt is null, give the seller pairingPage.guideUrl to install the extension, then pairingPage.url to pair it. Call this tool again after they finish. Input parameters: - `workspaceId` (string, required): The workspace to read extension pairing for, from spellbook_workspaces_list. Required. ### `spellbook_channel_pricing_profiles` (~125 tokens) Read which saved pricing profile each marketplace channel prices from in one workspace. Automatic eBay listing and repricing price every card from the eBay channel's profile, so read this before a seller turns on automatic eBay listing. Returns the workspace's saved profiles by name with the active one marked, and for each channel the pinned profile, the profile automation actually prices from, and which SpellBook jobs read that pin. It reports the setting and recommends no profile. Input parameters: - `workspaceId` (string, required): The workspace to read, from spellbook_workspaces_list. Required. ### `spellbook_channel_pricing_profile_set` (~222 tokens) Pin one of the workspace's saved pricing profiles to one marketplace channel, or clear the pin with a null profileId. It is the same choice as the channel defaults card on the SpellBook pricing settings page. Read spellbook_channel_pricing_profiles first, and pin only a profile the seller named. Returns the full channel pricing read after the write, plus the channel's previous pin, so the change can be reverted by setting the previous value back. It changes which profile prices future automatic listings and reprices; it edits no profile and publishes nothing. Input parameters: - `channel` (string, required): The marketplace channel to pin: ebay, tcg (TCGplayer), manapool, shopify or cardtrader. Required. - `profileId` (required): A profileId from spellbook_channel_pricing_profiles, or null to clear the pin so the channel falls back to the workspace's active profile. Required. - `workspaceId` (string, required): The workspace to change, from spellbook_workspaces_list. The signed-in customer must own it. Required. ### `spellbook_sku_price_controls` (~161 tokens) Read one seller SKU's exact minimum floor and exact-price lock by Scryfall printing, finish and condition. Returns saved values, update timestamps, the all-marketplaces lock scope, missing evidence and limitations. Call it before changing a floor or lock, then read it again after a write. Input parameters: - `condition` (string, required): The exact physical condition for the SKU. Required. - `finish` (string, required): The exact finish for the SKU: nf, fo, etched, surge, rh, 1e or 1e-fo. Required. - `scryfallId` (string, required): The exact Scryfall printing id. Required. - `workspaceId` (string, required): The workspace the seller selected. The caller must own it. Required. ### `spellbook_sku_minimum_floor_set` (~173 tokens) Set one seller SKU's minimum sell price and return its saved readback. Identify the exact printing, finish and condition. A floor raises computed prices up to the floor; it does not make an exact ask and does not trigger repricing. Input parameters: - `condition` (string, required): The exact physical condition for the SKU. Required. - `finish` (string, required): The exact finish for the SKU: nf, fo, etched, surge, rh, 1e or 1e-fo. Required. - `minSellPriceUsd` (number, required): The seller-selected minimum sell price in USD. Required. - `scryfallId` (string, required): The exact Scryfall printing id. Required. - `workspaceId` (string, required): The workspace the seller selected. The caller must own it. Required. ### `spellbook_sku_minimum_floor_clear` (~137 tokens) Clear one seller SKU's minimum floor by exact printing, finish and condition. Returns the saved controls readback, so the agent can confirm the floor is null. Input parameters: - `condition` (string, required): The exact physical condition for the SKU. Required. - `finish` (string, required): The exact finish for the SKU: nf, fo, etched, surge, rh, 1e or 1e-fo. Required. - `scryfallId` (string, required): The exact Scryfall printing id. Required. - `workspaceId` (string, required): The workspace the seller selected. The caller must own it. Required. ### `spellbook_sku_exact_price_lock_set` (~166 tokens) Set one seller SKU's exact price lock and return its saved readback. Identify the exact printing, finish and condition. The lock scope is all marketplaces and no marketplace publish or repricing is triggered by this tool. Input parameters: - `condition` (string, required): The exact physical condition for the SKU. Required. - `finish` (string, required): The exact finish for the SKU: nf, fo, etched, surge, rh, 1e or 1e-fo. Required. - `priceUsd` (number, required): The seller-selected exact locked price in USD. Required. - `scryfallId` (string, required): The exact Scryfall printing id. Required. - `workspaceId` (string, required): The workspace the seller selected. The caller must own it. Required. ### `spellbook_sku_exact_price_lock_clear` (~138 tokens) Clear one seller SKU's exact-price lock by exact printing, finish and condition. Returns the saved controls readback, so the agent can confirm the lock is null. Input parameters: - `condition` (string, required): The exact physical condition for the SKU. Required. - `finish` (string, required): The exact finish for the SKU: nf, fo, etched, surge, rh, 1e or 1e-fo. Required. - `scryfallId` (string, required): The exact Scryfall printing id. Required. - `workspaceId` (string, required): The workspace the seller selected. The caller must own it. Required. ### `spellbook_pricing_profiles_list` (~68 tokens) List seller-created pricing profiles by name and active state, plus the resulting active profile id. Call it before choosing a profile for detail, update, activation or channel pinning. Input parameters: - `workspaceId` (string, required): The workspace the seller selected. The caller must own it. Required. ### `spellbook_pricing_profile_get` (~94 tokens) Read one seller pricing profile's approved multiplier, floor, dollar offset, ordinary value bands and channel overrides, then report its active state and channel pin facts. Call it before an update or activation. Input parameters: - `profileId` (string, required): The exact saved pricing profile id from a prior pricing profile read. Required. - `workspaceId` (string, required): The workspace the seller selected. The caller must own it. Required. ### `spellbook_pricing_profile_create` (~151 tokens) Create one seller pricing profile from a seller-selected name and explicit baseline multiplier and floor, with optional dollar offset and ordinary channel overrides. Returns saved profile and pin readback. It does not activate, publish or reprice. Input parameters: - `channelOverrides` (object): Seller-approved overrides by marketplace channel. Omit when none are needed. - `floor` (number, required): Explicit baseline floor in USD. Required. - `marketOffsetUsd` (number): Optional baseline dollar adjustment. - `mult` (number, required): Explicit baseline market multiplier. Required. - `name` (string, required): Seller-selected profile name. Required. - `workspaceId` (string, required): The workspace the seller selected. The caller must own it. Required. ### `spellbook_pricing_profile_update` (~159 tokens) Update selected seller-approved fields on one pricing profile and return saved profile and pin readback. Omitted fields stay unchanged. Null clears only a selected channel override field. Input parameters: - `channelOverrides` (object): Selected per-channel overrides; omit to preserve all, null to clear one channel entry. - `floor` (number): Replacement baseline floor in USD. - `marketOffsetUsd` (number): Replacement baseline dollar adjustment. - `mult` (number): Replacement baseline multiplier. - `name` (string): Replacement seller-selected profile name. - `profileId` (string, required): The exact saved pricing profile id from a prior pricing profile read. Required. - `workspaceId` (string, required): The workspace the seller selected. The caller must own it. Required. ### `spellbook_pricing_profile_activate` (~92 tokens) Activate one seller-selected pricing profile and return its saved readback plus resulting active profile id. The first profile remains inactive until this explicit call. Activation does not publish or trigger repricing. Input parameters: - `profileId` (string, required): The exact saved pricing profile id from a prior pricing profile read. Required. - `workspaceId` (string, required): The workspace the seller selected. The caller must own it. Required. ### `spellbook_pricing_amendments_list` (~114 tokens) List bounded seller pricing amendments with selector, marketplace scope, absolute multiplier, protected band, enabled state, priority and timestamps. It returns a cursor plus missing evidence and limitations. Input parameters: - `cursor` (string): Opaque cursor from a prior list response. Omit for the first page. - `limit` (integer): Maximum amendments to return, 1 to 100. Omit for 50. - `workspaceId` (string, required): The workspace the seller selected. The caller must own it. Required. ### `spellbook_pricing_amendment_get` (~90 tokens) Read one saved pricing amendment's selector, marketplace scope, absolute multiplier, optional protected band, enabled state, priority and timestamps. Use it before an update, preview or delete. Input parameters: - `amendmentId` (string, required): The exact saved amendment id from a prior amendment read. Required. - `workspaceId` (string, required): The workspace the seller selected. The caller must own it. Required. ### `spellbook_pricing_amendment_create` (~176 tokens) Create one bounded seller pricing amendment from a selector, marketplace scope, absolute multiplier and optional protected band. Returns saved readback. The server retains its adaptive default internally; this tool has no demand behavior input and does not publish or reprice. Input parameters: - `enabled` (boolean): Whether the amendment starts enabled. Omit for enabled. - `marketplaceScope` (required): Marketplace scope. Required. - `mode` (object, required): Absolute multiplier and optional protected band. Required. - `name` (string, required): Seller-selected amendment name. Required. - `notes`: Optional seller notes; null clears notes. - `priority` (integer, required): Seller-selected precedence priority. Required. - `selector` (object, required): The bounded inventory selector. Required. - `workspaceId` (string, required): The workspace the seller selected. The caller must own it. Required. ### `spellbook_pricing_amendment_update` (~167 tokens) Update selected fields on one saved pricing amendment and return saved readback. Omitted fields preserve existing values, including server-owned demand behavior. The update does not publish or trigger repricing. Input parameters: - `amendmentId` (string, required): The exact saved amendment id from a prior amendment read. Required. - `enabled` (boolean): Replacement enabled state. - `marketplaceScope`: Replacement marketplace scope. - `mode` (object): Replacement absolute multiplier and optional protected band. - `name` (string): Replacement seller-selected name. - `notes`: Replacement notes; null clears notes. - `priority` (integer): Replacement precedence priority. - `selector` (object): Replacement bounded selector. - `workspaceId` (string, required): The workspace the seller selected. The caller must own it. Required. ### `spellbook_pricing_amendment_delete` (~84 tokens) Delete one saved pricing amendment by exact amendment id and return a deletion readback. This changes saved configuration and does not publish or trigger repricing. Input parameters: - `amendmentId` (string, required): The exact saved amendment id from a prior amendment read. Required. - `workspaceId` (string, required): The workspace the seller selected. The caller must own it. Required. ### `spellbook_pricing_amendment_preview` (~123 tokens) Start or run a bounded preview for one saved amendment or one seller-selected draft. Returns customer-safe per-channel impact counts and status evidence, with no raw inventory rows or demand telemetry. Preview is not publication. Input parameters: - `amendment` (object): Draft amendment; provide this or amendmentId. - `amendmentId` (string): The exact saved amendment id from a prior amendment read. Required. - `channels` (array): Optional marketplace channels to preview. - `workspaceId` (string, required): The workspace the seller selected. The caller must own it. Required. ### `spellbook_pricing_amendment_preview_status` (~98 tokens) Read one pricing amendment preview job's bounded status and per-channel customer impacts. It returns counts, quantities, delta totals, timestamps, missing evidence and limitations. It never publishes or reprices. Input parameters: - `jobId` (string, required): The preview job id returned by spellbook_pricing_amendment_preview. Required. - `workspaceId` (string, required): The workspace the seller selected. The caller must own it. Required. ### `spellbook_workspaces_list` (~68 tokens) List the authenticated customer's existing workspaces with workspaceId, name, membership role and status. Owner, admin and member roles are preserved. Discover a workspace before selecting it; this read creates no workspace or trial. When the list is empty, spellbook_workspace_create makes the first one. ### `spellbook_workspace_create` (~94 tokens) Create the signed-in seller's first workspace when spellbook_workspaces_list comes back empty. It makes a workspace the seller owns. A signup Store trial starts only when its flag is on and the free Seller cap is off for this workspace. When the seller already has a workspace, it returns the first one and creates nothing, so a repeated call is safe. It returns workspaceId, name, role, and created. ### `spellbook_store_setup_status` (~166 tokens) Read one workspace's store setup facts and the seller-activation projection: which setup facts are true (physical inventory imported, stock confirmed, channel activated, return address saved, first order synced, extension paired, ManaPool connected), the current phase, the single nextAction, and its blockers. It also returns ebayConnected, and each channel's automation mode with a plain meaning sentence in ebayAutomation.meaning and manapoolAutomation.meaning, so an agent can tell the seller which marketplaces are connected and which list cards automatically. Baseline quantity remains marketplace evidence, not physical stock. It applies to any question about where a seller stands in setup, or what remains before their store can sell. Input parameters: - `workspaceId` (string, required): The workspace to read setup status for. Required. ### `spellbook_ebay_automation_set` (~233 tokens) Set one workspace's eBay automation mode, which decides whether SpellBook lists and reprices on eBay by itself. This changes public eBay behavior for the workspace, only the workspace owner may call it, and the agent sets a mode only after the seller chose one. The modes: manual: Nothing lists or reprices on eBay automatically; you publish from the eBay preview when you choose. auto_list: New eligible inventory lists on eBay automatically at your saved eBay pricing strategy's prices, and existing eBay listings are never repriced automatically. full: New inventory lists on eBay automatically, and existing eBay listings reprice from your saved eBay pricing strategy. With variation packages on (the default), new inventory groups into one listing per set with each card as a variation; with them off, each card lists on its own. Input parameters: - `mode` (string, required): The mode the seller chose: manual, auto_list, or full. Required. - `workspaceId` (string, required): The workspace whose eBay automation mode to set. The caller must own it. Required. ### `spellbook_manapool_automation_set` (~263 tokens) Set one workspace's ManaPool automation mode, which decides whether SpellBook creates, updates, delists, and reprices ManaPool listings by itself. This changes public ManaPool behavior for the workspace, only the workspace owner may call it, and the agent sets a mode only after the seller chose one. The modes: manual: SpellBook creates no ManaPool listings and sends no quantity changes or delistings, but it keeps repricing your existing ManaPool listings. assisted: SpellBook plans the ManaPool listing changes it would make but sends none of them and offers no approval step, and it keeps repricing your existing ManaPool listings. automatic: SpellBook creates, updates, and delists ManaPool listings as your inventory changes and reprices existing ones, once the workspace passed a ManaPool test publish and while ManaPool reports the account ready to sell. paused: SpellBook sends nothing to ManaPool: no new listings, no quantity changes, no delistings, and no repricing. Input parameters: - `mode` (string, required): The mode the seller chose: manual, assisted, automatic, or paused. Required. - `workspaceId` (string, required): The workspace whose ManaPool automation mode to set. The caller must own it. Required. ### `spellbook_return_address` (~78 tokens) Read the signed-in seller's saved return address: the name, street line, and city-state-ZIP line SpellBook prints on shipping labels, plus the other settings stored beside it. profile is null when the seller never saved one. Read it before spellbook_return_address_save, so the agent can show the seller what is on file. ### `spellbook_return_address_save` (~244 tokens) Save the return address the seller gave, which SpellBook prints on shipping labels and which store setup counts as returnAddressSaved. It replaces only the name, street, and city-state-ZIP lines and keeps every other saved setting. Only the workspace owner may call it, and only with an address the seller stated. A seller who does not want to share a home address in chat can type it on a SpellBook page, which the limits name, so offer that page before you ask for the address. Call spellbook_store_setup_status afterwards to confirm returnAddressSaved. Input parameters: - `cityStateZip` (string, required): One line with city, state, and ZIP code, such as Springfield, IL 62701, exactly as the seller gave it. Required. - `name` (string, required): The name on the return address, such as the seller's name or store name, exactly as the seller gave it. Required. - `street` (string, required): One street line, such as a house number and street, exactly as the seller gave it. Required. - `workspaceId` (string, required): The workspace whose setup this address completes. The caller must own it. Required. ### `spellbook_workspace_plan_status` (~161 tokens) Read one workspace's Store plan and Seller cap status: tier, stored plan status, trialExpiresAt, sellerToolsUnlocked, and freeSellerCap's scope, observedAt, sellerSource, activeUnitCount, countIsExact, cap, headroom, overCap, enforcementEnabled, claimed, and transitionGraceExpiresAt. A free owner must explicitly choose one workspace with spellbook_free_seller_workspace_claim before free Seller access begins. Read this status again after the claim. For a seller-requested paid Store plan, use spellbook_seller_plan_checkout_link when the stored plan is not active, then re-read this status. Input parameters: - `workspaceId` (string, required): The workspace to read the plan for. The caller must be a member. Required. ### `spellbook_free_seller_workspace_claim` (~136 tokens) Claim the one permanent free Seller allowance for a workspace the seller explicitly chose. Ask the seller to name the workspace from spellbook_workspaces_list and confirm that choice before this call. Never pick a workspace from list order or an active-workspace default. The signed-in person must own it. Re-read spellbook_workspace_plan_status after this call for the actual allowance and entitlement. Input parameters: - `choiceConfirmed` (boolean, required): True only after the seller explicitly confirms this workspace choice in the current conversation. Required. - `workspaceId` (string, required): The exact workspaceId the seller explicitly selected for their one free Seller workspace. Required. ### `spellbook_seller_plan_checkout_link` (~171 tokens) Get a Stripe checkout page for a seller-chosen Store plan on one workspace. Read spellbook_workspace_plan_status first. Skip checkout when sellerToolsUnlocked comes from an active Store plan, person-level Pro, or admin access. When freeSellerCap.sellerSource is free_cap, offer paid Store access only if the seller chooses it to bypass the free active-unit cap. Give the returned URL to the seller to enter payment details at Stripe, then re-read plan status to observe its stored state; that status cannot prove payment settled. It refuses a workspace whose plan is already active. Input parameters: - `interval` (string, required): The billing interval the seller chose: monthly or annual. Required. - `workspaceId` (string, required): The workspace the Store plan is for. The caller must own it. Required. ### `spellbook_entitlements` (~29 tokens) Read this authenticated customer's current Investor and Seller access. Flags are not proof of payment or enrollment. ### `spellbook_singles_buylist_spreads` (~346 tokens) Investigate Magic TCGplayer NM acquisitions against Card Kingdom buylist payout scenarios. Preserve source timestamps, finish, stock, grading scenarios, and queue degradation; spread is not net realized profit. Input parameters: - `cursor` (string): Opaque cursor from a previous page, passed back unchanged. Omit to start from the first page. - `isFoil` (boolean): True restricts to foil rows, false to non-foil. Omit to include both. - `limit` (integer): Maximum rows to return, 1 to 200. Omit for the default of 20. - `minCkQty` (integer): Minimum Card Kingdom buylist quantity a row must have. Omit for no floor. - `minSpreadPct` (number): Minimum spread percentage a row must have to be returned. Omit for no floor. - `minSpreadUsd` (number): Minimum spread in USD a row must have to be returned. Omit for no floor. - `minTotalPotentialProfitUsd` (number): Minimum total potential profit in USD a row must have to be returned. Omit for no floor. - `minValue` (number): Minimum acquisition value in USD a row must have to be returned. Omit for no floor. - `rarity` (string): Exact rarity to filter to. Omit to include every rarity. - `sets` (string): Exact set code(s) to filter to. Omit to include every set. - `sortBy` (string): Which field ranks the rows. Omit for the default order. - `sortDir` (string): Sort direction for sortBy. Omit for the default direction. ### `spellbook_japanese_sourcing` (~379 tokens) Read Japanese-market supply offers for the Japanese promo cohort from Mercari, Hareruya and Bigweb, each beside the comparison prices SpellBook holds on TCGplayer, ManaPool and eBay. Every row carries one exact supply offer with its printing identity, the merchant's own condition grade and its normalized equivalent, the native unit price with its currency, quantity, availability, and its own observedAt and staleAfter stamps; then every comparison reference with its kind and typed comparisonBasis, a deterministic selectedReference with the selectionRule that chose it, and a providerStatus row for each comparison source explaining why it is ok, missing, stale, failed or incompatible. grossDifferenceUsd is an asking spread and is null unless printing, language, finish, unit basis and condition equivalence all hold; netDifferenceUsd stays null because this read carries no costs. Rows come back in source order, and nothing is ranked, scored or recommended for you. Input parameters: - `buySources` (array): Which Japanese supply sources to read. Omit for all three. Narrow it only when the person named a merchant; a source you drop is simply absent, never proven empty. - `comparisonSources` (array): Which comparison price sources to attach to each offer. Omit for all three. Each requested source returns a providerStatus row even when it has no reference. - `cursor` (string): The opaque nextCursor from the previous page, passed back unchanged. It is signed by the server; never construct, decode or edit one. - `includeStale` (boolean): True also returns offers whose staleAfter has already passed, with their stamps intact. Default false returns only offers still inside their freshness window. - `limit` (integer): Maximum offers in this page. Page with cursor to traverse the rest; a short page is not proof the cohort ended. ### `spellbook_japan_source_scan` (~172 tokens) Read the latest bounded Japan sourcing cohort: English Riftbound cards sold in Japan (riftbound-jp), Japanese-language Magic promos (japanese-promos), or exact-language Pokemon cards (pokemon, pokemon-en, pokemon-ja). Returns every tracked target's actual observation time, verified offers, exact reference prices with price kinds, unresolved listing URLs and recorded blockers, source status and freshness policy. publishedAt is snapshot publication time and does not refresh an older target. Sold and unrelated observations are not purchase candidates. Missing costs remain unknown; no buying recommendation or profit is provided. Input parameters: - `cohortId` (string): English Riftbound cards sold in Japan, Japanese-language Magic promos, or one exact-language Pokemon cohort. Use pokemon-en or pokemon-ja to keep source observations isolated. Defaults to riftbound-j… ### `spellbook_riftbound_japanese_comparables` (~124 tokens) Read physical Japanese-language Riftbound observations, separate from English cards sold in Japan. Returns exact SKU language, finish, condition, price kind, source links, actual price capture timestamps, freshness policy, per-source eligible denominators and nullable acceptance rates. ManaPool is unsupported_game. Distinguishes unavailable, stale, completed eligible zero and accepted evidence. TCG Market is a provider statistic, not a completed sale or guaranteed resale proceeds. asOf is query time. An empty result never implies a market price of zero or proves market-wide coverage. ### `spellbook_pricing_decision_explain` (~165 tokens) Read one seller listing's recorded pricing evidence for one marketplace. Returns the last recorded ask, market anchor, safe price effect, decision timestamps and whether the evidence is observed, recorded, projected or missing. Call this when a seller asks what price was saved or what evidence supports it, then use the seller controls tools when they want to inspect or change a setting. It never claims a remote marketplace ask is current without a remote observation timestamp. Input parameters: - `channel` (string, required): Which marketplace listing to explain. Required, because the same copy carries a different ask per channel. - `itemId` (string, required): The exact Item id, from a prior spellbook_inventory_items_query. Required. - `workspaceId` (string, required): The workspace that owns the listing. Required. ### `spellbook_credit_conversion_candidates` (~66 tokens) Read CK credit-funded sealed acquisition candidates ranked by opening EV. Cash/credit opening multiples do NOT rank intact sealed resale cash profit. Credit is restricted purchasing power. Input parameters: - `boxType` (string): Restrict candidates to one box type. Defaults to all, covering every type. ### `spellbook_credit_resale_assessment` (~192 tokens) Assess up to ten exact sealed product lines as a CK-credit-funded intact resale scenario. Supply starting credit and explicit cost assumptions; omit unknown costs. EV judgment is permitted. Returns labeled cash estimates, credit allocation, capacity and missing evidence. Research only, no purchase. Input parameters: - `costs` (object, required): Explicit cost assumptions for the resale scenario. Every field is optional; omit any cost that is unknown rather than guessing zero. - `exitChannel` (string, required): Name of the channel the intact resale would exit through. Required, up to 30 characters. - `lines` (array, required): Up to 10 exact sealed product lines to assess. Required, at least one. - `maxSnapshotAgeHours` (number, required): Maximum age in hours of price snapshots this assessment may use. Required. - `startingCreditUsd` (number, required): Card Kingdom store credit available to spend in this scenario, in USD. Required. ### `spellbook_sealed_deal_board` (~175 tokens) Read today's Sealed Deal Board rows with their age stamps, so you can answer "what is new since " without diffing the whole board yourself. Each row carries firstSeenAt (the day the board first carried that product), firstSeenBasis, priceBeatenAt with priceBeatenFromUsd (the day its cheapest price fell at least 5% and at least $10 under its own trailing 30-day low), and recencyEventAt, which is the later of the two. Optionally filter to one game; omit game to include every available game. Filter and compare these yourself; this returns facts and ranks nothing by recency. Input parameters: - `game` (string): Restrict rows to one game: mtg, riftbound or pokemon. Omit to include all games. ### `spellbook_sealed_product_search` (~322 tokens) Search sealed catalog identity and available offer records by a third-party listing's name, set and product type. catalogMatches covers boxes, packs and cases, returning exact product IDs, kind, physicalLanguage, form, contentsState, intact marketPrice with marketPriceObservedAt, and coverage (priceState, modelState, evState, totalEv, missingInput) independently of EV or offer availability. products returns deal rows with modeled totalEV and offer evidence when supported. Call spellbook_sealed_product_offers with a deal row's tcgProductId for retail references, available sold comparables and freshness. Public or entitled response is chosen by the backend; teaser and gating fields limit deal coverage. Input parameters: - `cursor` (string): Opaque cursor from a previous page, passed back unchanged. Omit to start from the first page. - `game` (string): Restrict results to one game: mtg, riftbound or pokemon. Omit for a mixed-game search. - `limit` (integer): Maximum rows to return, 1 to 100. Omit for the default of 20. - `productType` (string): Exact sealed product type to filter to. Omit to include every type. - `q` (string): Free-text product name search. Omit to browse without a name filter. - `region` (string): Restrict results to one currency region: US, CA or All. Omit for all. - `store` (string): Exact store name to filter to. Omit to include every store. ### `spellbook_sealed_product_offers` (~144 tokens) Inspect one exact TCG product ID's sealed offers and their source times. Each offer includes the exact store listing, its local price, the crawl's actionability observation, and the selected exit comparable with its evidence date. Checkout shipping, tax, and available quantity are explicitly unknown until checked with the store. Opening EV is not intact resale value. Empty title or null computedAt can mean missing snapshot. Input parameters: - `region` (string): Restrict retail references to one currency region: US, CA or All. Omit for all. - `tcgProductId` (string, required): Exact TCGplayer product id for this sealed product, digits only. Required. ### `spellbook_support_report` (~222 tokens) Store one minimal customer failure report. Generate a fresh ULID submissionId and reuse the exact request for retries within 24 hours. Reports are accepted only after storage. No raw diagnostics, conversations, secrets, or customer identity. Available to authenticated Free customers, including paid-access denials. A report reaches no person. Input parameters: - `clientVersion` (string, required): This client's semantic version, e.g. "1.2.3". Required. - `correlationId` (string): A correlation id from a prior response, if one was returned. Omit when none was given. - `errorCode` (string, required): Which closed error code best matches the failure. Required. - `occurredAt` (string, required): ISO-8601 timestamp of when the failure occurred. Required. - `operation` (string, required): Which SpellBook operation failed. Required, one of the closed operation names. - `submissionId` (string, required): A fresh ULID for this report. Reuse the exact same value and request body to retry within 24 hours; a new value starts a new report. ### `spellbook_agent_gap_report` (~365 tokens) After finishing the task, ask the person whether they consent to send one bounded capability-gap report to SpellBook. Call this tool only after an explicit yes for this report and set consentToShare to true. Use it when a needed field was null, you guessed a number because no tool supplied it, you wished a tool existed, or you abandoned a path. It returns no task data and fixes nothing. Generate a fresh ULID submissionId and reuse the exact request for retries within 24 hours. Field paths must be real SpellBook response paths. There is no free-text field: never describe the card, price, customer, or conversation. Input parameters: - `abandonReason` (string): Why the task was abandoned. Required only when outcome is abandoned. - `consentToShare` (boolean, required): True only after the person explicitly agreed to send this specific report. - `guessedFields` (array): Real SpellBook response field paths the agent had to guess a value for because no tool supplied one. Omit when nothing was guessed. - `neededFields` (array, required): Real SpellBook response field paths that were null or missing when needed. Omit or send empty when none applied. - `occurredAt` (string, required): ISO-8601 timestamp of when the gap occurred. Required. - `outcome` (string, required): What happened to the task: closed enum. Required. - `submissionId` (string, required): A fresh ULID for this report. Reuse the exact same value and request body to retry within 24 hours; a new value starts a new report. - `taskCategory` (string, required): Which closed task category this gap happened in. Required. - `wishedCapability` (object): A capability the agent wished existed and did not find. Omit when no capability was missing. ### `spellbook_customer_evidence_record` (~105 tokens) Append one immutable consideration, realized outcome, or work observation to the selected workspace. A receipt proves storage only. It does not verify source providers, inspected card truth, profit, or completed work. Input parameters: - `request` (required): The evidence envelope: a submissionId, occurredAt, a kind (realized_outcome, consideration or work_observation), and that kind's payload. Required. - `workspaceId` (string, required): The workspace to append this evidence to. Required. ### `spellbook_customer_evidence_get` (~87 tokens) Read one private immutable evidence receipt from the selected workspace. Returned observations and source fields remain submitted evidence. Input parameters: - `receiptId` (string, required): The receipt's receiptId from the record response, or the submissionId you chose when recording it. Either one resolves the same receipt inside the workspace. Required. - `workspaceId` (string, required): The workspace the receipt belongs to. Required. ### `spellbook_customer_evidence_list` (~188 tokens) List a bounded page of private evidence receipts from the selected workspace, optionally filtered by receipt kind, parent, or item ID. Input parameters: - `cursor` (string): Opaque cursor from a previous page, passed back unchanged. Omit to start from the first page. - `itemId` (string): Restrict the list to receipts referencing this exact item id. Omit for no item filter. - `kind` (string): Restrict the list to one receipt kind. Omit to include every kind. - `limit` (integer): Maximum receipts to return, 1 to 50. Omit for the default of 20. - `parentReceiptId` (string): Restrict the list to receipts linked to this parent receipt's receiptId from the record response. Omit for no parent filter. - `workspaceId` (string, required): The workspace to list receipts from. Required. ### `spellbook_inventory_items_query` (~158 tokens) Read one private page of existing physical Items matching exact bins or printing, finish and condition groups. Input parameters: - `bins` (array): Restrict results to these exact bin names, up to 25. Omit to search every bin. - `cursor` (string): Opaque cursor from a previous page, passed back unchanged. Omit to start from the first page. - `groups` (array): Restrict results to these exact printing, finish and condition groups, up to 25. Omit to skip this filter. - `limit` (integer): Maximum Items to return, 1 to 1000. Omit for the default of 20. - `workspaceId` (string, required): The workspace to read Items from. Required. ### `spellbook_inventory_holdings` (~198 tokens) Read one page of what a seller's own workspace holds, aggregated to one row per printing, finish and condition. Each row carries card name, set code, collector number, variant label, quantity owned, quantity available, the current unit market price and the resolved minimum-sell floor with the tier that produced it. Call this to answer what is in the collection, what it is worth, what is already listed versus still available, and which rows sit near their floor. It reports holdings; it does not rank them, pick what to sell, or state a listing's live marketplace status. Input parameters: - `cursor` (string): Opaque cursor from a previous page, passed back unchanged. Omit to start from the first page. - `limit` (integer): Maximum rows to return, 1 to 500. Omit for the default of 100. - `workspaceId` (string, required): The workspace to read holdings for. Required. ### `spellbook_sealed_portfolio` (~124 tokens) Read a seller's sealed boxes, packs and cases, with quantity, cost basis, market price, heldQuantity, openIntentQuantity, acquisitionSource and workspaceId for each purchase entry. Call this to reconcile physical sealed stock before listing or opening it. Recovered components can have unknown allocated cost basis and an approximate recovery date. A hold differs from an actual opening; call spellbook_sealed_portfolio_hold_set to correct a holding mark. Input parameters: - `workspaceId` (string, required): The workspace whose sealed positions to read, from spellbook_workspaces_list. Required. ### `spellbook_sealed_portfolio_record` (~302 tokens) Record one sealed acquisition the customer has already made, so the buy can be scored later against what it cost. Names the catalog boxId, the quantity, the unit price paid as a decimal string, and the acquisition date. Use it after a seller buys a box, pack or case found through spellbook_sealed_deal_board. It records a purchase that happened; it never buys anything and moves no money. Input parameters: - `acquisitionChannel` (string): Where the product was bought, such as a store name or marketplace. Optional, and it is stored as the customer stated it. - `acquisitionDate` (string, required): The date the buy happened, as YYYY-MM-DD. Required, and it sets which day's modeled EV the purchase is scored against. - `acquisitionPrice` (string, required): Price paid PER UNIT in USD, as a decimal string such as "114.99". Not the order total. Required. - `boxId` (string, required): Exact sealed catalog product id, from spellbook_sealed_product_search or spellbook_sealed_deal_board. An id the catalog does not hold is refused. Required. - `quantity` (integer, required): How many units of this product were bought. Required. - `salesTaxUsd` (string): Sales tax paid on the whole order in USD, as a decimal string. Optional. - `workspaceId` (string, required): The workspace to record this sealed buy in, from spellbook_workspaces_list. Required. ### `spellbook_sealed_portfolio_hold_set` (~134 tokens) Mark every still-sealed unit in one purchase entry as a long-term hold, or release that hold. A hold keeps physical inventory and market value but excludes those units from sealed listing. Read spellbook_sealed_portfolio before and after to check heldQuantity. Input parameters: - `entryId` (string, required): The exact purchase entryId from spellbook_sealed_portfolio, including the SEALED# prefix when present. Required. - `hold` (boolean, required): True excludes live units from listing; false makes them eligible again. Required. - `workspaceId` (string, required): The workspace that owns the entry. Required. ### `spellbook_sealed_portfolio_open_intent_clear` (~106 tokens) Clear the plan to open every still-sealed unit in one purchase entry. Call this when spellbook_sealed_portfolio shows openIntentQuantity for stock the seller intends to keep sealed; then read the portfolio again. This preserves the units and any holding mark. Input parameters: - `entryId` (string, required): The exact purchase entryId from spellbook_sealed_portfolio. Required. - `workspaceId` (string, required): The workspace that owns the entry. Required. ### `spellbook_sealed_portfolio_receive` (~102 tokens) Mark one incoming sealed display purchase entry as physically received after the seller confirms its arrival. Call spellbook_sealed_portfolio first to identify the incoming entry, then read it again to confirm owned status. This preserves its holding mark and opening plan. Input parameters: - `entryId` (string, required): The exact incoming display entryId from spellbook_sealed_portfolio. Required. - `workspaceId` (string, required): The workspace that owns the incoming entry. Required. ### `spellbook_sealed_portfolio_open` (~139 tokens) Record that an exact quantity of sealed boxes or packs has already been physically opened. Use the entryId from spellbook_sealed_portfolio and read the returned counts; the same entryId and quantity safely replay after a timeout. The historical opening date stays unknown, and recordedAt means when SpellBook recorded the status. No singles are created. Input parameters: - `entryId` (string, required): The exact sealed purchase entryId from spellbook_sealed_portfolio. Required. - `quantity` (integer, required): Exact count of units already opened from this entry. Required. - `workspaceId` (string, required): The workspace that owns the sealed entry. Required. ### `spellbook_sealed_recovered_component_record` (~243 tokens) Record sealed packs recovered from already-opened products, without adding a new purchase or guessing their share of the parent product's cost. This bounded intake accepts the confirmed Aetherdrift Box Topper and Modern Horizons 3 Collector Booster Sample Pack products. It records zero incremental cash, keeps parent basis unallocated, marks recovery timing approximate, and returns persisted-row readback. Reuse requestId unchanged after a timeout. Input parameters: - `hold` (boolean, required): True marks these units as held from listing; false leaves them unheld. Required. - `productId` (string, required): Exact TCG sealed pack product id: 618698 for Aetherdrift Box Topper, or 579917 for Modern Horizons 3 Collector Booster Sample Pack. Required. - `quantity` (integer, required): How many recovered units to record: up to 11 for product 618698, or exactly 1 for product 579917. Required. - `requestId` (string, required): Stable idempotency key for this recovery entry. Reuse this exact value to retry. Required. - `workspaceId` (string, required): The workspace to add recovered sealed components to. Required. ### `spellbook_sealed_intent_defaults` (~74 tokens) Read a workspace's future sealed-purchase open or hold defaults. Call this before changing how extension-captured products are classified. These rows affect future captures only; spellbook_sealed_portfolio reports marks already on physical units. Input parameters: - `workspaceId` (string, required): The workspace whose defaults to read. Required. ### `spellbook_sealed_intent_default_set` (~138 tokens) Set a workspace rule for future extension-captured sealed purchases. An exact game and productType wins over wildcard rules. Read spellbook_sealed_intent_defaults before and after; existing units keep their current marks. Input parameters: - `game` (string, required): The game to match, or * for every game. Required. - `intent` (string, required): The capture intent: open or hold. This does not physically open or list stock. Required. - `productType` (string, required): The sealed product type to match, or * for every type. Required. - `workspaceId` (string, required): The workspace whose rule to set. Required. ### `spellbook_sale_record` (~478 tokens) Record a sale of cards the seller's workspace holds that happened outside SpellBook's order intake, such as a local game store buy, an in-person trade, or a one-off marketplace sale SpellBook did not import. Names the exact printing, finish, condition, quantity and unit sale price. It marks that many copies sold, oldest first, closes the position so spellbook_inventory_holdings and spellbook_buy_outcome_report reflect it, and pulls those copies off every marketplace where they are still live. Returns the sold item ids and the holding after the sale. Input parameters: - `binId` (string): Sell only copies from this bin. Omit to take the oldest copies from any bin. - `buyerName` (string): The buyer's name as the customer stated it. Optional. - `condition` (string, required): The condition of the sold copies, matching how they were recorded. Required. - `finish` (string, required): The finish of the sold copies, matching the finish recorded in inventory. Pokemon supports nf (Normal/Unlimited), fo (Holofoil), rh (Reverse Holofoil), 1e (1st Edition) and 1e-fo (1st Edition Holofoi… - `marketplace` (string, required): Where the sale happened. in_person covers a store buy or a trade. Required. - `note` (string): A free-text note stored with the sale. Optional. - `occurredAt` (string): When the sale happened, as an ISO-8601 timestamp. Omit to use the time of this call. - `orderId` (string, required): A stable reference for this sale, such as the buyer's order number or a receipt number. Required. A second call with the same orderId for the same printing, finish and condition returns the first sal… - `productId` (string, required): The exact printing id, from spellbook_printings_search or spellbook_inventory_holdings. Required; a wrong id sells the wrong card. - `qty` (integer, required): How many copies sold. Required. - `salePricePerUnit` (number, required): Price received PER COPY in USD, before any fee. Not the order total. Required. - `workspaceId` (string, required): The workspace that held the sold copies, from spellbook_workspaces_list. Required. ### `spellbook_inventory_location_confirm` (~168 tokens) Save up to 25 existing Item bin locations after comparing each expectedBins observation with current state. A fresh Item query is required before and after the save. Input parameters: - `expectedBins` (array, required): One prior-bin observation per itemId in itemIds, up to 25. Required; a missing or changed Item rejects the whole batch. - `itemIds` (array, required): Up to 25 exact Item ids to move, from a prior spellbook_inventory_items_query. Required, at least one. - `sourceSignalId` (string): An id linking this confirmation to its originating signal. Omit when there is none. - `toBin` (string, required): The bin these Items are being confirmed into. Required. - `workspaceId` (string, required): The workspace the Items belong to. Required. ## Diagnostics Captured diagnostic sections: TLS, DNSSEC, Authorisation, Transports. The full working is on the page: https://verifymcp.io/servers/com-spellbook-finance-mcp/api#diagnostics ## Score history - 2026-09-28: 66 - 2026-09-27: 66 - 2026-09-26: 65 - 2026-09-25: 65 ## Common questions ### What is the SpellBook Finance MCP server? SpellBook Finance is an MCP server listed in the public MCP registry as com.spellbook-finance/mcp. Magic: The Gathering card prices, movers, arbitrage, sealed box EV, and seller inventory tools. This page covers its hosted endpoint (https://api.spellbook-finance.com/mcp). ### Is the SpellBook Finance MCP server safe to use? SpellBook Finance scores 66 out of 100 on VerifyMCP. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm. ### What tools does the SpellBook Finance MCP server expose? SpellBook Finance exposes 100 tools: spellbook_guides, spellbook_skill, spellbook_pull_work, spellbook_buy_targets, spellbook_buy_outcome_report, and 95 more. Their descriptions and schemas cost roughly 18,207 tokens of context every time the server is loaded. ### Does the SpellBook Finance MCP server require authentication? No. We connected to SpellBook Finance without credentials and it answered, so anything it exposes is reachable by anyone who knows the address. ### Is the SpellBook Finance MCP server still maintained? SpellBook Finance is still listed as active in the MCP registry. We last reached this channel on 28 September 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced. ## Links - Remote endpoint: https://api.spellbook-finance.com/mcp - Website: https://spellbook-finance.com/agent/mcp.md - Changelog RSS feed: https://verifymcp.io/servers/com-spellbook-finance-mcp/api.xml - Changelog JSON feed: https://verifymcp.io/servers/com-spellbook-finance-mcp/api.json - HTML version of this page: https://verifymcp.io/servers/com-spellbook-finance-mcp/api