@cruxible-ai
基本情報
Hard, verifiable state layer for AI agents
以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。
{
"mcpServers": {
"cruxible": {
"command": "cruxible-mcp",
"env": {
"CRUXIBLE_MODE": "admin"
}
}
}
}ツール一覧はまだありません
この掲載には表示できるツール一覧がまだありません。サーバーにツールがないという意味ではありません。対応機能はプロジェクトのドキュメントをご確認ください。
ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。
Cruxible Core is a deterministic, typed state layer for AI agents and human teams to operate on verifiable, governed domain state. It models entity and relationship types, write rules, and recurring workflows in a Terraform-like config, and the runtime enforces them. The core contains no LLM or hidden API calls—it works with any agent harness and pins exports from real systems as artifacts.
Install with pip install cruxible, then start a local daemon with CRUXIBLE_SERVER_STATE_DIR="$HOME/.cruxible/sandbox" cruxible server start. Initialize a domain with cruxible init --kit <name>, connect an agent session with cruxible context connect, run deterministic workflows with cruxible run and cruxible apply, propose governed claims with cruxible propose, and resolve them with cruxible group resolve. Queries return receipted results via cruxible query run.
Cruxible Core models the domain as typed, governed state instead of retrieving similar chunks of text. Claims carry provenance and review state, multi-hop traversal follows exact relationships, and results are exact and repeatable across reads. Markdown and retrieval remain appropriate for drafts and exploration, but Cruxible Core certifies durable truth outside the model.
Cruxible Core runs on Python 3.11+. It is installed via `pip install cruxible` and launched as a local HTTP server. No LLM, external API, or database is required—state is local to a state directory (`CRUXIBLE_SERVER_STATE_DIR`), and sandbox mode uses a built-in `operator` identity.
When `CRUXIBLE_SERVER_AUTH=true` is set, identity is enabled: a bootstrap credential is claimed, and each agent receives its own token attributing every write. The server listens by default on `http://127.0.0.1:8100`. Token-based permission tiers and hardening are documented in the Runtime Auth And Agent Roles guide.
Health by Dr Brad works out a health plan from your own results: which screening is due at your age, and evidence-based suggestions, each with the reason and the published guideline or study behind it. Your record is one
A good sequence diagram can solve almost any communication problem in engineering. Volley lets your agent write them as plain text, with the real request and response payloads on every arrow. It checks its own syntax wi
Yuppsy MCP: build and publish real estate websites from your AI assistant. Yuppsy lets your AI assistant create live websites for real estate agents and individual property listings. You don't need a designer or any code
DeepMark helps teachers deliver rapid, consistent marking with meaningful feedback for every student — in a fraction of the time. What once took a week, now takes one free period.
A curated list of references for MLOps
コメント