cassandra-predict
Cassandra prediction hub: one MCP door over govcon forecasts (recompete, contractor signals).
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": {
"cassandra-predict": {
"url": "https://predict.dropwatchhq.com/mcp"
}
}
}遠端端點
https://predict.dropwatchhq.com/mcpstreamable-http它能做什麼
工具清單
工具（2）
⚪predict(domain, entity)
Cassandra unified predictor. From leading signals in public data, returns what is about to happen to an entity — weeks early — across four verticals. domain selects the vertical: 'recall' (consumer product / vehicle / drug / device recall risk from complaint velocity), 'bio' (biotech NIH→trial→SEC catalyst chain), 'recompete' (federal contract recompete-winner shift), 'ma' (government contractor acquisition target). entity is the name to look up (e.g. 'Tesla Model Y', 'Moderna', a contractor or agency name). Returns a leading-signal prediction with lead time — NOT a guarantee, and NOT a claim of any official action.
輸入結構描述
{
"type": "object",
"properties": {
"domain": {
"type": "string",
"enum": [
"recall",
"bio",
"recompete",
"ma"
],
"description": "Prediction vertical to route to."
},
"entity": {
"type": "string",
"description": "Entity name / ticker / id to predict on."
}
},
"required": [
"domain",
"entity"
]
}🟢list_domains
List Cassandra prediction verticals available via predict(domain, entity), with what each predicts and its data sources.
輸入結構描述
{
"type": "object",
"properties": {}
}社群
證據