MCP GSC AI Agent Carrier Class Fleet
BPC Sustainable Carrier Agents - Low Token Zero Inference
我該用這個嗎
品質與安全性
發現項目（5）
- HIGHTool poisoning patterns detected
- MEDIUMTool description contains URL to non-standard domain在 ask_elyssah 中
- LOWTool 'list_fleets' description lacks action verb在 list_fleets 中
- LOWTool 'resolve_jurisdiction' description lacks action verb在 resolve_jurisdiction 中
- LOWTool 'how_to_transact' description lacks action verb在 how_to_transact 中
根據工具定義與協定合規性的自動化分析。
上下文成本
這是每次將伺服器的工具載入模型上下文時所消耗的約略 token 數量。數量越高，可用於其他工作的注意力就越少。
安裝
一鍵安裝
將以下內容加入你的 `claude_desktop_config.json` 檔案：
{
"mcpServers": {
"gsc-fleet-mcp": {
"url": "https://mcp.gsc-fleet.ai/mcp"
}
}
}遠端端點
https://mcp.gsc-fleet.ai/mcpstreamable-http它能做什麼
工具清單
工具（11）
🟢list_fleets
The three live fleet roots — Global, APAC, LATAM — with regions, slot ranges, and entry points. Headline: 32,597 LIVE AI Agents — the 32,500+ AI Agent program: 19,597 Carriers on the fleet roots and 14 sovereign nodes, 3,000 on the APP surfaces, 10,000 AI Orderability (AIO) Agents on the ten gsc-marketplace.ai hosts.
輸入結構描述
{
"type": "object",
"properties": {},
"$schema": "http://json-schema.org/draft-07/schema#"
}🟢get_fleet(fleet)
Full record for one fleet: root, region, slot range, index, catalog, Card URL pattern.
輸入結構描述
{
"type": "object",
"properties": {
"fleet": {
"type": "string",
"enum": [
"global",
"apac",
"latam"
],
"description": "Fleet key"
}
},
"required": [
"fleet"
],
"$schema": "http://json-schema.org/draft-07/schema#"
}🟢resolve_carrier(fleet, slot)
(fleet, slot) → the canonical Card URL. Validates the slot against the fleet's real range; the Card itself is the source of truth — fetch it to confirm the Carrier.
輸入結構描述
{
"type": "object",
"properties": {
"fleet": {
"type": "string",
"enum": [
"global",
"apac",
"latam"
],
"description": "Fleet key"
},
"slot": {
"type": "string",
"pattern": "^\\d{4}$",
"description": "Four-digit slot number, e.g. 0001"
}
},
"required": [
"fleet",
"slot"
],
"$schema": "http://json-schema.org/draft-07/schema#"
}🟡find_carriers(node, category)
SM-ECO-10060 member code → the serving surface: a resident sovereign endpoint where one exists, otherwise the fleet root covering that member's region. Category and capability addressing ship with the sovereign Carrier set and will answer here as they go live.
輸入結構描述
{
"type": "object",
"properties": {
"node": {
"type": "string",
"maxLength": 4,
"description": "SM-ECO-10060 member code, e.g. UK, KR, BR"
},
"category": {
"type": "string",
"maxLength": 64,
"description": "Optional SPARKS category (answers as the sovereign set ships)"
}
},
"required": [
"node"
],
"$schema": "http://json-schema.org/draft-07/schema#"
}🟢list_jurisdictions
Where GSC Carriers are resident — eighteen jurisdictions, each with its serving Azure region. Residency is the checkable claim.
輸入結構描述
{
"type": "object",
"properties": {},
"$schema": "http://json-schema.org/draft-07/schema#"
}⚪resolve_jurisdiction(code)
Member code → residency status, serving endpoint, and Azure region.
輸入結構描述
{
"type": "object",
"properties": {
"code": {
"type": "string",
"maxLength": 4,
"description": "Member code, e.g. CH"
}
},
"required": [
"code"
],
"$schema": "http://json-schema.org/draft-07/schema#"
}🟢get_residency_receipt
How to verify a GSC surface's residency claim on the wire — which headers to read and what they prove.
輸入結構描述
{
"type": "object",
"properties": {},
"$schema": "http://json-schema.org/draft-07/schema#"
}🟢get_fleet_stats
Canon numbers: 32,597 LIVE AI Agents (32,500+ AI Agent program: 22,597 Carriers + 10,000 AIO Agents) across eighteen resident jurisdictions. Monthly transaction volume is published at gsc-radar.ai — the source of record.
輸入結構描述
{
"type": "object",
"properties": {},
"$schema": "http://json-schema.org/draft-07/schema#"
}🟢get_carrier_doctrine
The Carrier AI Agent class and the sell-side positioning — prefabricated, static, zero inference; buyer-side compute stays the buyer's.
輸入結構描述
{
"type": "object",
"properties": {},
"$schema": "http://json-schema.org/draft-07/schema#"
}⚪how_to_transact
Which GSC surface an agent uses for data, discovery, transaction, and standards — with the endpoint for each.
輸入結構描述
{
"type": "object",
"properties": {},
"$schema": "http://json-schema.org/draft-07/schema#"
}⚪ask_elyssah(q, prev)
Ask GSC's AI assistant elyssah a question about GSC, its AI Agents for Beauty & Personal Care brands, the numbers of record, the AI investment market per the analysts, management or coverage. Returns the same JSON the human desk gets: answer {q, a, id, anchor, door} served byte for byte from GSC's record (no model in the answer path), shelf (up to 3 related pairs on record), suggested links, and sig — the estate keyring signature (kid gsc-cards-2026-08) over {id, a, ts}, verifiable at https://gsc-registry.ai/keyring.json. Pass prev = the id of the last answer for one-turn context. Not investment advice.
輸入結構描述
{
"type": "object",
"properties": {
"q": {
"type": "string",
"minLength": 1,
"maxLength": 300,
"description": "The question, plain words, up to 300 characters"
},
"prev": {
"type": "string",
"maxLength": 40,
"description": "id of the last answer shown (one-turn context, ranking only)"
}
},
"required": [
"q"
],
"$schema": "http://json-schema.org/draft-07/schema#"
}社群
證據