Kanonik
ai.kanonik由 kanonik.ai 發布官方發布者v1.0.0更新於 Oct 24, 2026
Governance runtime for compliance: verified, human-approved writes to a tamper-evident record.
概覽
Kanonik 是遠端治理執行環境，讓 AI 在人工核准、防竄改紀錄的前提下起草、評估並記錄合規工作。
- 可以做什麼
- 依已載入的合規框架（如 ISO 27001:2022、SOC 2、GDPR、NIST CSF 2.0、HIPAA）起草與修訂政策及程序，並產出適用性聲明或該框架的等效文件。評估內部控制並記錄每項判斷背後的證據。找出框架要求與工作區實際狀況之間的落差。建立查核人員可獨立驗證的稽核軌跡。每次寫入都經伺服器端 Verifier 檢查，並在進入紀錄前由人工核准。
- 適用情境
- 適用於需要結果經得起檢視的合規工作：起草或修訂政策、評估控制、進行落差分析，或為公認框架整理稽核證據與適用性聲明。適合想借助 AI、但要求每次變更都經人工核准並留下防竄改紀錄的團隊。
- 執行需求
- 遠端 MCP 端點 app.kanonik.ai，透過 streamable HTTP 存取，無需本機套件或程序。需要 Kanonik 帳號，首次使用時透過瀏覽器完成 OAuth 2.1 登入。合規框架由工作區提供。需要能連線至 Kanonik 端點與驗證服務的網路。
安裝
在 SourceWeft 中
- 開啟 儀表板中的 Kanonik，將其新增到工作區。
- 連接 所需的 憑證。它們會被私密儲存，絕不會在公開頁面上展示。
- 為需要使用其工具的對話啟用該服務。
Web executable，透過 Streamable HTTP。 遠端服務在工作區中設定後即可從網頁執行環境執行。
交給代理程式安裝
把這段提示詞貼上到 Claude Code、Codex、Cursor 等代理程式，或 SourceWeft 對話中。代理程式會讀取該伺服器的安裝指南，先向你展示發布者、端點和登入方式，經你同意後再把它新增到你的客戶端。金鑰不會寫進任何命令。
閱讀 https://sourceweft.com/mcp/ai.kanonik%2Fkanonik/install.md，並按其中說明把這個 MCP 伺服器接入我的客戶端。修改任何設定之前，先告訴我它的發布者、端點和登入方式，等我確認。不要把 API 金鑰寫進命令或我會提交的檔案；問我金鑰應從哪裡取得。其他 MCP 客戶端
把它新增到你客戶端的 mcpServers 設定中。 客戶端會在首次連接時完成 OAuth 登入。
{
"mcpServers": {
"kanonik": {
"type": "http",
"url": "https://app.kanonik.ai/mcp"
}
}
}README
Kanonik
Governance runtime for compliance work, connected to Grok over MCP.
Kanonik gives your AI a typed compliance model to work in, a tamper-evident record of everything it does, a server-side Verifier that checks its work, and a human-approval gate in front of every write. Your AI does the reasoning. Kanonik makes the result defensible.
What you can do with it
Ask your AI to work on your compliance program in plain language:
- Draft and revise policies and procedures against a loaded framework
- Assess internal controls and record the evidence behind each judgement
- Find the gaps between what a framework requires and what you actually have
- Build an audit trail an assessor can verify independently
- Produce a Statement of Applicability, or the equivalent artifact for your framework
The framework in scope comes from your workspace. Kanonik runs ISO 27001:2022, SOC 2, GDPR, NIST CSF 2.0 and HIPAA today. Nothing in the plugin is specific to any one of them.
Installing
Then pick Kanonik. Or from the CLI:
On first use your AI connects to https://app.kanonik.ai/mcp and a browser
window opens for sign-in. You will need a Kanonik account; sign up at
kanonik.ai.
What this plugin contains
One file that tells Grok where the Kanonik MCP server lives:
That is the whole plugin. There is no bundled code, no install script, no hook, and no local process. Everything runs server-side at Kanonik.
Permissions and data
- Network. One endpoint,
https://app.kanonik.ai/mcp. Nothing else. - Authentication. OAuth 2.1 with Dynamic Client Registration and PKCE,
against
https://auth.kanonik.ai. Tokens are held by your Grok client. The plugin holds no credential and ships no secret. - Scopes.
kanonik:readto read your compliance record,kanonik:proposeto draft changes,kanonik:committo submit them for approval. - Writes. No change reaches your record without a signed single-use approval token and a human clicking to approve it. Your AI cannot approve its own work.
- Your other systems. Kanonik never holds credentials to your source systems. If your AI reads from GitHub, Jira, or your identity provider, it does so through its own connections, not through Kanonik.
Links
- Site: kanonik.ai
- Documentation: kanonik.ai/docs
- Contact: kanonik.ai/contact
License
See LICENSE. The Kanonik service itself is proprietary; this repository contains only the connection manifest and documentation for it.
來源：README.md，提交 6ea4498
工具
0版本歷史
1- v1.0.0最新Sep 16, 2026
評分
已滿足適用檢查項的 85%
該評分衡量此收錄條目的完整程度，依據註冊中心條目、端點與公開來源計算得出。它不是對伺服器本身的評價，也不表示伺服器安全或目前可用：可連通性是帶日期的檢查結果，評分也不包含安裝量或使用者評價。
- 近 30 天內可連通必需25 分
- 已宣告認證方式必需10 分
- 已列出工具15 分
請讓伺服器在無憑證時回應 tools/list，或發布可在沙箱中無憑證啟動的軟體包。
- 描述必需15 分
- 圖示10 分
- 文件15 分
- 授權條款10 分