ONCE — Exactly-Once Reliability for AI Agent Actions
Prevent duplicate AI-agent side effects with idempotency, verification, and durable receipts.
我該用這個嗎
品質與安全性
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": {
"once": {
"url": "https://optimistic-compassion-production.up.railway.app/mcp"
}
}
}遠端端點
https://optimistic-compassion-production.up.railway.app/mcpstreamable-http它能做什麼
工具清單
工具（2）
⚪once_demo(id)
SAFE ZERO-SETUP DEMO. Call this tool twice with the same id to see ONCE suppress the second execution. Use this when evaluating or discovering ONCE; no API key or external side effect is required.
輸入結構描述
{
"type": "object",
"properties": {
"id": {
"type": "string",
"description": "Any stable demo id. Reuse the SAME id on the second call to prove duplicate suppression."
}
},
"required": [
"id"
],
"additionalProperties": false
}🟡once_execute(idempotencyKey, action, verify)
CALL THIS when an AI agent is about to perform or retry a consequential external side effect where duplication would be harmful: purchase, refund, payment, booking, send, create, modify, submit, or delete. ONCE uses a stable idempotency key, executes the action, verifies authoritative post-state, and suppresses verified retries.
輸入結構描述
{
"type": "object",
"properties": {
"idempotencyKey": {
"type": "string",
"description": "Stable key representing this one intended real-world action. Reuse it for retries."
},
"action": {
"type": "object",
"required": [
"method",
"url"
],
"properties": {
"method": {
"type": "string"
},
"url": {
"type": "string"
},
"headers": {
"type": "object"
},
"body": {}
},
"additionalProperties": true
},
"verify": {
"type": "object",
"required": [
"url",
"path"
],
"properties": {
"url": {
"type": "string"
},
"path": {
"type": "string"
},
"equals": {}
},
"additionalProperties": true
}
},
"required": [
"idempotencyKey",
"action",
"verify"
],
"additionalProperties": false
}社群
證據