High-performance, token-optimized Model Context Protocol (MCP) server for Flipkart Seller Hub v3 APIs.
Five tools with complete parameter schemas and descriptions, but significant gaps in output documentation, error handling, and composition clarity. All tools follow verb_noun naming (flipkart_manage_*), but the 'manage' verb is overloaded across 5 distinct domains (shipments, inventory, returns, ads, financials), creating cognitive load. Parameter descriptions are present but generic; many lack constraints (enums, ranges, formats). No documented output schemas, pagination guidance, or error recovery paths. Tools are action-multiplexed (single tool with 'action' parameter doing 4-7 different things), violating single-responsibility principle and forcing LLMs to reason about action strings rather than tool selection.
Manage Flipkart advertising campaigns (PLA / PCA / Smart ROI), budgets, and bidding. Actions: - 'list_campaigns': View active or paused ad campaigns. - 'get_metrics': Fetch campaign performance (impressions, clicks, spend, orders, ROAS). - 'update_budget': Update daily budget (INR). - 'update_bid': Modify keyword or product bid prices. Output format: Returns token-optimized Markdown table ('summary') or JSON ('full').
Manage Flipkart seller bank settlements, payouts, and financial/tax report generation. Actions: - 'get_settlements': View payout summaries, bank settlement status, UTR numbers. - 'request_report': Trigger generation of sales, tax, inventory, or returns reports. - 'download_report': Check status and retrieve download link for generated report. Output format: Returns token-optimized Markdown table ('summary') or JSON ('full').
Manage Flipkart product listings, inventory levels, pricing, and catalog CRUD. Actions: - 'search': Search catalog listings by SKU or FSN. - 'get_details': Fetch full listing specifications and current pricing/stock. - 'update_stock': Modify available inventory counts across warehouses. - 'update_price': Update MRP and Selling Price (INR). - 'create_listing': Create a new catalog listing (Title, SKU, Selling Price, Brand). - 'update_listing': Update listing title, brand, or custom attributes. - 'delete_listing': Deactivate or archive a listing by SKU. Output format: Returns token-optimized Markdown table ('summary') or JSON ('full').
Action-multiplexed tools violate single-responsibility principle. Each tool accepts 4-7 distinct actions (search, inspect, mark_rtd, get_label, cancel for shipments; search, get_details, update_stock, update_price, create_listing, update_listing, delete_listing for inventory). LLMs must reason about action strings rather than selecting the right tool. This increases hallucination risk and prevents proper tool composition.
No output schemas documented. Tool descriptions mention 'Returns token-optimized Markdown table (summary) or JSON (full)' but do not specify field names, types, or structure. LLMs cannot plan downstream calls or extract required IDs (shipment_id, sku, return_id, campaign_id, report_id) without seeing actual response format.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 54 | 2026-07-28+ | v2 |
Manage Flipkart customer returns, RTO tracking, and Seller Protection Fund (SPF) claims. Actions: - 'list_returns': List customer returns, RTOs, or replacements. - 'track_rto': Track return shipment delivery status and courier location. - 'approve_return': Approve customer return request. - 'spf_claim': File SPF claim for damaged, wrong, or missing return items. Output format: Returns token-optimized Markdown table ('summary') or JSON ('full').
Manage Flipkart seller orders & shipments. Actions: - 'search': List pending, packed, RTD, or dispatched orders (filter by state/order_id). - 'inspect': Retrieve detailed order items, SLA deadlines, customer pincode. - 'mark_rtd': Transition orders to Ready to Dispatch state. - 'get_label': Fetch printable shipping label & tax invoice PDF URLs. - 'cancel': Cancel order shipment with specified reason. Output format: Returns token-optimized Markdown table ('summary') or JSON ('full').
Missing enum constraints on action parameters. 'action' accepts free-form strings (search, inspect, mark_rtd, etc.) without formal enum declaration. LLMs will hallucinate invalid actions like 'list_shipments' or 'get_shipment_status'. Enums are self-documenting and prevent invalid invocations.
No error handling guidance. Tool descriptions do not explain what happens on failure (invalid action, missing shipment_id, API rate limit, authentication failure). No recovery paths provided. LLMs cannot self-correct or know whether to retry.
Parameter descriptions lack format/constraint details. 'date_from' and 'date_to' in financials lack format specification (YYYY-MM-DD assumed but not stated). 'stock' in inventory has no range (0-999999?). 'daily_budget' in ads has no currency or min/max. LLMs will guess and pass invalid values.