Enforces commit-time policy checks for AI agent writes with ALLOW, DENY, or STEP_UP decisions and a tamper-evident audit log.
別名 AvenueDAdmin/pywrit, writ-mcp
- スター
- 2
- ツール
- 15
- ダウンロード数（30日間）
- —
- 最新バージョン
- 0.1.2
- 最終コミット
- 2026年11月13日
- ライセンス
- MIT
Writ MCPとは
WritはMCPサーバーです。 Enforces commit-time policy checks for AI agent writes with ALLOW, DENY, or STEP_UP decisions and a tamper-evident audit log. 15個のツールを提供しており、writ_check, writ_verify_token, writ_grantなどがあります。
プロジェクトのREADMEから生成された概要
Writ MCP server について
Writ provides point-of-action control for AI agents by enforcing policy checks before any consequential write operation occurs. This MCP server uses a stdio transport to return ALLOW, DENY, or STEP_UP decisions, ensuring that agents operate within defined boundaries. It is designed for developers using MCP-compatible clients like Claude Desktop or Claude Code, requiring a WRIT_API_KEY for full functionality.
The system implements a check-before-write loop where the agent calls writ_check to receive a 90-second authToken. This token is then validated via writ_verify_token immediately before the action is executed. For higher security, the STEP_UP decision requires a human sponsor to approve the request through writ_grant. Every decision generates a tamper-evident audit receipt accessible via writ_receipts, providing a clear history of agent actions and authorizations.
できること
- 1Verify agent write permissionsAsk the agent to check if it has permission to perform a specific write action before it proceeds.
- 2Secure sensitive operations with tokensUse a time-bound token to ensure an authorized write operation is still valid at the moment of execution.
- 3Require human intervention for risky tasksTrigger a manual approval process when an agent attempts an action that requires a step-up decision.
- 4Audit agent decision historyRetrieve a detailed trail of all previous authorization decisions and outcomes.
プロジェクトの README から作成されています。すべてのユースケースは、その引用と照らし合わせて確認されています。
Writ MCPのインストール方法
フォーマット検証日 2026年11月10日Claude CodeにWrit MCPをインストール
$ claude mcp add --env WRIT_API_KEY=<YOUR_WRIT_API_KEY> --env WRIT_SPONSOR_TOKEN=<YOUR_WRIT_SPONSOR_TOKEN> writ -- uvx writ-mcpCursorにWrit MCPをインストール
~/.cursor/mcp.json
{
"mcpServers": {
"writ": {
"command": "uvx",
"args": [
"writ-mcp"
],
"env": {
"WRIT_API_KEY": "<YOUR_WRIT_API_KEY>",
"WRIT_SPONSOR_TOKEN": "<YOUR_WRIT_SPONSOR_TOKEN>"
}
}
}
}VS CodeにWrit MCPをインストール
.vscode/mcp.json
{
"servers": {
"writ": {
"type": "stdio",
"command": "uvx",
"args": [
"writ-mcp"
],
"env": {
"WRIT_API_KEY": "<YOUR_WRIT_API_KEY>",
"WRIT_SPONSOR_TOKEN": "<YOUR_WRIT_SPONSOR_TOKEN>"
}
}
}
}Claude DesktopにWrit MCPをインストール
claude_desktop_config.json
{
"mcpServers": {
"writ": {
"command": "uvx",
"args": [
"writ-mcp"
],
"env": {
"WRIT_API_KEY": "<YOUR_WRIT_API_KEY>",
"WRIT_SPONSOR_TOKEN": "<YOUR_WRIT_SPONSOR_TOKEN>"
}
}
}
}CodexにWrit MCPをインストール
~/.codex/config.toml
[mcp_servers.writ] command = "uvx" args = ["writ-mcp"] [mcp_servers.writ.env] WRIT_API_KEY = "<YOUR_WRIT_API_KEY>" WRIT_SPONSOR_TOKEN = "<YOUR_WRIT_SPONSOR_TOKEN>"
Gemini CLIにWrit MCPをインストール
~/.gemini/settings.json
{
"mcpServers": {
"writ": {
"command": "uvx",
"args": [
"writ-mcp"
],
"env": {
"WRIT_API_KEY": "<YOUR_WRIT_API_KEY>",
"WRIT_SPONSOR_TOKEN": "<YOUR_WRIT_SPONSOR_TOKEN>"
}
}
}
}Writ MCPツール 15
READMEから検出writ_checkwrit_verify_tokenwrit_grantwrit_receiptswrit_policywrit_sandboxwrit_github_installationswrit_github_set_thresholdwrit_github_toggle_pr_commentswrit_manifestswrit_explain_receiptwrit_tenant_notify
| writ_check | — | |
| writ_verify_token | — | |
| writ_grant | — | |
| writ_receipts | — | |
| writ_policy | — | |
| writ_sandbox | — | |
| writ_github_installations | — | |
| writ_github_set_threshold | — | |
| writ_github_toggle_pr_comments | — | |
| writ_manifests | — | |
| writ_explain_receipt | — | |
| writ_tenant_notify | — |
15個のツールをすべて表示ツールを折りたたむ
writ_principalswrit_verbswrit_keys
| writ_principals | — | |
| writ_verbs | — | |
| writ_keys | — |
設定
WRIT_API_KEY必須シークレットWrit API key (get one free at https://withwrit.com)
WRIT_SPONSOR_TOKENシークレットWrit sponsor token, only required for sponsor tools (grant, revoke, reinstate)
WRIT_BASE_URLOptional custom Writ API base URL (default: https://api.withwrit.com)
| 名前 | 説明 | 種類 |
|---|---|---|
| WRIT_API_KEY | Writ API key (get one free at https://withwrit.com) | 必須シークレット |
| WRIT_SPONSOR_TOKEN | Writ sponsor token, only required for sponsor tools (grant, revoke, reinstate) | シークレット |
| WRIT_BASE_URL | Optional custom Writ API base URL (default: https://api.withwrit.com) |
自動チェック
実行日 2026年11月12日詳細レポート →自動チェックに合格しました。これはセキュリティ監査ではありません。コードを読み、サーバーに必要最小限のアクセス権を与えてください。 チェックの仕組み
- ソースリポジトリがリンクされており、アクセス可能合格
- オープンソースライセンス (OSI)合格
- 過去1年以内にコミットまたはリリースあり合格
- パッケージ名が有名なパッケージに酷似していない合格
- リモートエンドポイントがHTTPSを使用未チェック
- 破壊的なツールがラベル付けされている合格
- モデルへの隠れた指示がない合格
- パッケージと依存関係に重大なアドバイザリがない合格
README
GitHubで見るwrit-mcp
The Writ gate as an MCP server. Give any MCP-compatible
agent point-of-action control: the agent calls Writ before a consequential
write, gets back ALLOW, DENY, or STEP_UP, and every outcome creates an
audit receipt.
Install
pip install writ-mcp
Requires Python 3.9+. Installs the writ-mcp command (stdio transport).
Configure
export WRIT_API_KEY="writ_..." # required; get one free: POST /v1/keys with an email
export WRIT_SPONSOR_TOKEN="..." # only for sponsor tools (grant, revoke, reinstate)
# export WRIT_BASE_URL="..." # default: https://api.withwrit.com
Add to your agent
Claude Code:
claude mcp add writ --env WRIT_API_KEY="$WRIT_API_KEY" -- writ-mcp
Claude Desktop (claude_desktop_config.json):
{
"mcpServers": {
"writ": {
"command": "writ-mcp",
"env": { "WRIT_API_KEY": "writ_..." }
}
}
}
Any MCP client (generic stdio config):
{
"command": "writ-mcp",
"env": {
"WRIT_API_KEY": "writ_...",
"WRIT_SPONSOR_TOKEN": "writ_sp_..."
}
}
Hermes (NousResearch/hermes-agent) catalog entry — once published, this
package is installable from the optional-mcps catalog:
hermes mcp install writ
The check-before-write loop
The tool descriptions teach the agent this flow, but the short version:
writ_check(sponsor_id, agent_id, verb, target, purpose)— before the write.ALLOW→ you get a 90-secondauthTokenbound to that exact write.writ_verify_token(auth_token, verb, target, purpose)— immediately before executing, to prove the authorization still matches what you're doing.DENY→ do not proceed.STEP_UP→ a human sponsor approves (viawrit_grantor the dashboard), then check again.
Tools
| Tool | Who | What |
|---|---|---|
writ_check |
agent | Decision + 90s purpose-bound token |
writ_verify_token |
agent | Point-of-action token verification |
writ_grant |
sponsor | Approve a STEP_UP (one-time grant) |
writ_revoke / writ_reinstate |
sponsor | Kill switch on/off for a principal |
writ_receipts |
agent | Audit trail of decisions |
writ_policy |
agent | Tenant verb policy |
writ_sandbox |
anyone | Free 90-second demo grant, no key needed |
writ_github_installations |
admin | List connected GitHub App installations |
writ_github_set_threshold |
admin | Set the check-run fail threshold per install |
writ_github_toggle_pr_comments |
admin | Enable/disable the PR-comment bot |
writ_manifests |
agent | List repos with risk profiles or fetch one manifest |
writ_explain_receipt |
agent | Explain a receipt in plain language |
writ_tenant_notify |
admin | Get or update step-up notification settings |
writ_principals |
admin | List the tenant's principals |
writ_verbs |
admin | List the tenant's verb vocabulary |
writ_keys |
admin | List API key metadata |
Try it with no key: writ_sandbox → writ_check with verb="demo_write".
Source
Public repo: withwrit/pywrit (mcp/
directory). License: MIT.
Directory listings: Glama ·
MCP Registry io.github.withwrit/writ.
Writ MCPのよくある質問
Writ MCPをClaude Codeに追加するには
ターミナルでclaude mcp add --env WRIT_API_KEY=<YOUR_WRIT_API_KEY> --env WRIT_SPONSOR_TOKEN=<YOUR_WRIT_SPONSOR_TOKEN> writ -- uvx writ-mcpを実行します。すべてのプロジェクトで利用するには--scope userを追加します。
Writ MCPをCursorに追加するには
~/.cursor/mcp.jsonを開き、上記のインストールセクションに示すCursor設定を使ってmcpServersの下にサーバーを追加します。
Writ MCPが提供するツールは
15個のツールを提供しており、writ_check, writ_verify_token, writ_grant, writ_receipts, writ_policyなどがあります。
How does the check-before-write loop work?
The agent calls writ_check to receive an ALLOW, DENY, or STEP_UP decision; an ALLOW provides a 90-second authToken which is then verified via writ_verify_token before execution.
Can I try Writ without an API key?
Yes, you can use the writ_sandbox tool for a free 90-second demo grant without needing a key.
What happens if a STEP_UP decision is returned?
A human sponsor must approve the action via writ_grant or the dashboard before the agent can proceed.