LucidShark
Local code quality and security scanner: lint, types, SAST, SCA, tests, coverage.
Descripción del creador (original en inglés)
Save and use 1Password secrets locally without exposing plaintext to AI agents.
Estos tres nunca se combinan en una sola puntuación: las opiniones, las mediciones y la popularidad son cosas distintas.
Todavía no escribimos un resumen sencillo para este. Abajo están los datos que leímos del registro.
Elige la app de IA que usas y copia la configuración.
Necesitas
¿Qué app usas?
Elige tu app arriba. La recordamos, así que la próxima vez el botón dirá “Instalar en …”.
Abre Claude Desktop → Settings → Developer → Edit Config y pega esto en claude_desktop_config.json
{
"mcpServers": {
"onepassword-agent": {
"command": "npx",
"args": [
"-y",
"onepassword-agent-mcp",
"mcp"
]
}
}
}Ejecuta esto en tu terminal
claude mcp add onepassword-agent -- npx -y onepassword-agent-mcp mcpPégalo en .cursor/mcp.json de tu proyecto, o en ~/.cursor/mcp.json
{
"mcpServers": {
"onepassword-agent": {
"command": "npx",
"args": [
"-y",
"onepassword-agent-mcp",
"mcp"
]
}
}
}Tu navegador te pide permiso para abrir la app; luego la app muestra la configuración y te pide confirmar. Los valores de las claves quedan vacíos para que los completes tú.
Pégalo en .vscode/mcp.json de tu proyecto
{
"servers": {
"onepassword-agent": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"onepassword-agent-mcp",
"mcp"
]
}
}
}Tu navegador te pide permiso para abrir la app; luego la app muestra la configuración y te pide confirmar. Los valores de las claves quedan vacíos para que los completes tú.
Este se ejecuta en tu computadora, y ChatGPT solo puede conectarse a servidores en internet. Elige otra app.
Tu elección se guarda en una cookie en este dispositivo; no necesitas cuenta.
Esta configuración se genera a partir de la ficha del creador en el registro. Si el README dice otra cosa, sigue el README.
Se ejecuta en tu computadora; la lista de herramientas aparece en tu app de IA después de instalarlo.
Revisamos información que cualquiera puede ver; no hemos leído el código. Las revisiones que no aplican no se cuentan. · Revisado el 8 nov 2026
El nombre en el registro se verificó con la cuenta de GitHub gambadio, dueña del repo
Todavía no hemos leído este repo en GitHub
Todavía no hemos leído este repo en GitHub
Este se ejecuta en tu computadora; no hay endpoint en internet
Se ejecuta en tu computadora y usa ajustes locales en su lugar
Se ejecuta en tu computadora; solo podemos leer listas de herramientas de endpoints en internet
Comparado con todos los servidores activos de esta categoría en el registro. Las estrellas solo cuentan cuando quien lo publica es dueño del repo.
Superó 1 de 1 revisiones básicas aplicables. No hemos leído el código. Antes de agregar una clave o un token, revisa qué acceso pide el servidor y usa una clave que tenga solo los permisos que necesitas.
El registro no indica ninguna clave obligatoria, aunque algunos servidores te piden iniciar sesión la primera vez que los usas.
Se ejecuta en tu computadora. Tu app de IA lo inicia cuando hace falta, y los datos no pasan por los servidores del creador (a menos que el propio servidor llame a un servicio en línea).
El repo de GitHub cambió por última vez –, y la última versión (0.4.2) llegó al registro hace 12 días.
Las apps compatibles con MCP, como Claude Desktop, Claude Code, Cursor y VS Code. También necesitas Node.js en tu computadora.
En la misma categoría: LucidShark, Threatist, Draugr
Servidores de la misma categoría hechos por otros creadores, ordenados por estrellas.
Cómo leer una tarjeta:★ Estrellas de GitHubiTres pilares: reseñas · revisiones · usoiActualizado = fecha en que se publicó la última versión en el registroi
Local code quality and security scanner: lint, types, SAST, SCA, tests, coverage.
Vulnerability records: CVE, GHSA, CISA KEV, EPSS and vendor fixes in one place. No key needed.
Security scanning for AI agents: SAST, SCA, secrets, IaC, DAST, ranked by real risk.
Copiado tal cual del repo del creador. PopNx MCP no lo ha revisado ni lo respalda. Se quitaron HTML, imágenes y scripts. Los enlaces llevan fuera de este sitio.
Let AI agents use and save 1Password secrets through a local MCP server, with a dashboard that puts you in control.
npm package · setup guide · user guide · security model · uninstall guide
Independent project. Not affiliated with or endorsed by 1Password. Every installation uses its own local settings and 1Password account.
*The current dashboard, rendered with synthetic example items. All screenshots in this README use demo data; no real account, password, or service-account token is shown. Open full-size Vault screenshot.*
You need Node.js 20+, the 1Password CLI (`op`), and an MCP client that supports local stdio servers. Authenticate using 1Password desktop integration or a scoped service account.
On macOS, install the prerequisites if needed:
brew install node 1password-cli
For desktop authorization, open 1Password → Settings → Developer and enable Integrate with 1Password CLI. Approve only clients you trust. Service-account setup is described under Unattended Access.
Install and run the guided installer:
npm install -g onepassword-agent-mcp
onepassword-agent-mcp install
The installer detects supported MCP clients and asks before updating their configuration. On macOS, it separately offers the optional visible menu-bar shortcut.
Start the dashboard:
onepassword-agent-mcp admin
Open http://127.0.0.1:7319. For a manual setup, use onepassword-agent-mcp doctor, then onepassword-agent-mcp setup all --apply.
Updating an existing installation? Run npm install -g onepassword-agent-mcp@latest, then restart the admin console and existing MCP clients so they use the new code. Your local settings, approvals, and encryption key are retained.
The sidebar separates the four everyday tasks:
| View | What you control |
|---|---|
| Vault | Create the agent vault, copy or move items, approve exact fields, and manage Allowed For Agents. |
| Permissions | Enable item creation, clipboard capture, and automatic approval; configure unattended access and the optional Mac menu-bar shortcut. |
| Profile | Add profile data, such as email or address, that agents may read directly. |
| Activity | Review recent local operations without secret values. |
Copy keeps the source item. Move removes it from the source vault after transfer. Copying an item into the agent vault does not approve it. An empty website list allows all websites, so add a domain when access should be limited.
The type filters—All Fields, Logins, API Keys, Credit Cards, Notes & SSH, and Other—make larger vaults easier to review. Sensitive fields such as card CVV and PIN are not suggested automatically during manual approval.
To transfer a newly saved credential back to another vault, open Review Details → Save this item to another vault. Copy To Vault keeps the agent-vault item; Move To Vault removes it from the agent vault and removes its local approvals. Use the 1Password app for items containing attachments, documents, or passkeys.
Open Permissions, choose the settings you want, then click Save Settings. All three options below are off on a fresh installation.
| Setting | Effect |
|---|---|
| Allow agents to save new items into MCPVAULT | Allows supported new items to be saved into the configured agent vault. |
| Save secrets from the clipboard | Allows metadata-only clipboard saves. Item creation must also be enabled. |
| Automatically approve items agents create | Approves supported fields after a new save is verified, so agents can use them immediately. |
*Demo configuration with the three opt-in permissions enabled and simulated saved service-account status. The token field is empty; the saved token is never displayed. Open full-size Permissions screenshot.*
View the responsive dashboard
*The same local dashboard at a narrow viewport. This screenshot does not imply remote access; the server binds to loopback by default.*
The Profile view is separate: values entered there are intentionally readable by agents through get_profile_data. Do not store passwords or tokens as profile values.
In Permissions, enable Allow agents to save new items into MCPVAULT and Save secrets from the clipboard, then click Save Settings. Optionally enable automatic approval if agents should be able to use new items immediately.
{
"title": "Example account", "requestId": "example-signup-unique-id", "category": "login", "url": "https://example.com", "username": "my-login" }
Use those arguments with save_secret_from_clipboard. Supported categories are login, password, api_credential, and secure_note. Text is preserved exactly, including spaces and newlines. The tool runs on the computer hosting this MCP server and reads that computer's clipboard; a remote browser's clipboard is a different clipboard.
The app encrypts a retry snapshot locally, usable for 15 minutes, passes plaintext only through the local CLI's stdin, and verifies the stored item. The response contains save and approval metadata, never the copied text; use the find/list tools for approved encrypted handles. Expired ciphertext is removed while the process runs or on the next capture after downtime; a metadata-only record prevents reusing an old request ID. Clipboard clearing rechecks that the value has not changed and is best effort; check the returned cleanup status. Clipboard managers may retain their own history.
Reuse the same requestId and metadata to retry. A retry uses the original encrypted capture even if the clipboard changes. An expired capture fails closed instead of capturing unrelated text; inspect the previous operation before intentionally starting a new request. Metadata such as titles, usernames, and URLs remains visible to the model: never put a secret in those fields. save_secret_item remains available for values already in context; it cannot undo that exposure.
Under Permissions, enable Automatically approve items agents create, then click Save Settings. This option defaults to off. Agent item creation must also be enabled.
When enabled, a verified save through save_secret_item or save_secret_from_clipboard automatically approves every supported field of the new item in the configured agent vault, including credit-card CVV and PIN. The agent can use the resulting encrypted handles immediately. Existing items and items copied or moved in through the admin console still need manual approval.
CLI controls, site scoping, and retry behavior
The CLI controls the same saved option:
onepassword-agent-mcp settings auto-approve on
onepassword-agent-mcp settings auto-approve off onepassword-agent-mcp settings --json
Changes apply to subsequent new saves without restarting the MCP server. Add --dry-run to preview a change. Turning the option off keeps existing approvals; disable or delete individual approvals in Allowed For Agents to revoke them. The automatic approval choice is recorded when each save starts. Retrying an older save or a save that opted out does not opt it into automatic approval, and retries do not reinstate revoked approvals.
Automatic approvals use allowedSites when provided, otherwise the saved item's website. An empty list or an item without a website allows all sites. approveForAgents: false skips approval for that save. With automatic approval off, approveForAgents: true retains the existing single-field approval behavior, using approveField when supplied and allowing all sites unless allowedSites is provided. With automatic approval on, approveField does not narrow the set of approved fields.
The Permissions → Unattended access card accepts a 1Password service-account token directly in the local dashboard. Create the token in 1Password's service-account settings with access only to your dedicated agent vault. Grant read_items, plus write_items for saves. In the dashboard, paste it into Service account token and click Save encrypted token. The input clears immediately. Do not send the token in chat.
The app encrypts the token locally in ~/.onepassword-mcp/access.json, never returns it to the dashboard or MCP client, and supplies it only to the CLI process environment. Changes apply to later CLI calls without restarting. Test connection checks vault visibility; it does not write a test item or certify write permissions. Turn off Use saved service account to switch back to desktop authentication, or choose Remove saved token to delete the local credential. Revoke the token in 1Password if you want to invalidate it everywhere. In-flight operations may finish after a toggle.
An externally configured OP_SERVICE_ACCOUNT_TOKEN takes precedence; the dashboard reports that mode, and removing a local token cannot disable the environmen
Solo se muestra el principio; lee el resto en el repo