Sgraal Memory Governance
@sgraal-ai
Sgraal Memory Governance について
Deterministic pre-action verdict gate for agent memory — USE/WARN/ASK/BLOCK before an agent acts on a stored memory. Governs your existing store, doesn't replace it.
基本情報
設定
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": {
"sgraal": {
"command": "npx",
"args": [
"@sgraal/mcp"
],
"env": {
"SGRAAL_API_KEY": "sg_demo_playground"
}
}
}
}ツール
ツールは検出されませんでした
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。
概要
What is Sgraal Memory Governance?
Sgraal Memory Governance is a deterministic preflight gate for AI-agent memory. Before an agent acts on a stored memory, it returns a USE / WARN / ASK / BLOCK verdict, catching poisoned or stale memory at the read boundary. It is non-LLM and deterministic: identical input yields identical verdict. It governs whatever memory store you already use, not another store.
How to use Sgraal Memory Governance?
A demo key is provided: sg_demo_playground — no signup needed. The README does not specify further install, configure, or invoke steps.
Key features of Sgraal Memory Governance
- Pre-action verdict on every memory read: USE / WARN / ASK / BLOCK
- Blocks unsafe memory before an agent acts on it
- Four deterministic detection layers: timestamp, identity, consensus, provenance
- Multi‑tenant containment limits poisoned memory propagation
- Tested against an internal adversarial corpus of 1,190+ cases
Use cases of Sgraal Memory Governance
- Preventing an agent from acting on poisoned or stale memories
- Adding a deterministic safety gate to any existing memory store
- Containing the spread of compromised memories across tenants
- Enforcing memory provenance and identity checks in multi‑agent systems
FAQ from Sgraal Memory Governance
What verdicts does Sgraal Memory Governance return?
It returns one of four deterministic verdicts: USE, WARN, ASK, or BLOCK.
Does it replace my existing memory store?
No. It governs whatever memory store you already use; it is not another store.
Is Sgraal Memory Governance LLM‑based?
No, it is deterministic and non‑LLM; identical input yields identical verdict.
How was it tested?
Tested against an internal adversarial corpus of over 1,190 cases; independent third‑party evaluation is in progress.
How can I try it?
Use the demo key sg_demo_playground — no signup required.
「メモリとナレッジ」の他のコンテンツ
MCP server for Obsidian
MarkusPfundsteinMCP server that interacts with Obsidian via the Obsidian rest API community plugin
Memory Bank MCP Server
alioshrA Model Context Protocol (MCP) server implementation for remote memory bank management, inspired by Cline Memory Bank.
Obsidian MCP Server
cyanheadsRead, write, search, and surgically edit Obsidian vault notes, tags, and frontmatter via MCP. STDIO or Streamable HTTP.
mcp-local-rag
nkapila6"primitive" RAG-like web search model context protocol (MCP) server that runs locally. ✨ no APIs ✨
Zettelkasten MCP Server
entanglrA Model Context Protocol (MCP) server that implements the Zettelkasten knowledge management methodology, allowing you to create, link, explore and synthesize atomic notes through Claude and other MCP-compatible clients.
コメント