topview-amazon-ops

Topview Amazon Ops

Public Amazon market / ASIN / keyword intelligence through the host's unified topview-mcp server (progressive disclosure; domain amazon).

Step 0 — Bootstrap

1. Confirm the host already has MCP server topview-mcp connected (look for use_amazon_data, get_tool_schema, call_tool).
2. If tools are missing, tell the user to add the server from repo mcp.example.json into ~/.cursor/mcp.json, reload MCP — then retry.
3. Do not run scripts/mcp_call.py, open SSE yourself, or use curl/raw HTTP.

Execution

Server: topview-mcp (unified). Never open SSE yourself.

Amazon data tools are deferred — they do not appear as top-level tools/list entries. Invoke via progressive disclosure:

1. Read the use_amazon_data description on tools/list (embedded Tools catalog). Calling use_amazon_data only to list tools is optional; call it once if the catalog says it is not loaded yet (Amazon proxy warm-up).
2. Pick a deferred tool name from the catalog or references/tool_routing.md.
3. get_tool_schema with { "name": "<deferred>", "domain": "amazon" }.
4. call_tool with { "name": "<deferred>", "domain": "amazon", "arguments": { ... } }.

Typical host tool ids: mcp__topview-mcp__get_tool_schema / mcp__topview-mcp__call_tool (Cursor may prefix user-).
Do not assume deferred names like market_research are direct top-level tools. Creative tools (topview_* / ta_*) are on the same server but belong to topview-generate — call those directly, never via call_tool.

UX Rules

1. Be concise; lead with findings, not RPC noise.
2. Ask for marketplace / category / ASIN / keyword when required by the tool schema.
3. Never invent private Seller Central Ads/order/profit numbers.
4. Match the user's language.
5. After analysis, if the user wants creatives, chain to topview-generate.

Source

Routing adapted from Marketing Studio Amazon expert skills (V032).