All servers / Security / 1Password MCP server
1Password MCP server
Security-first MCP server for 1Password — vault/item tools, prompts, resources, and op_run secret injection (MCP 2026-07-28)
Tested by ConnectATDScore 90 out of 100. Tested 2026-11-19 on package version 5.1.1. Package: @takescake/1password-mcp · source
Score breakdown
| Runs | 18 / 20 | |
|---|---|---|
| Tool descriptions | 20 / 20 | |
| Package safety | 20 / 20 | |
| Maintenance | 20 / 20 | |
| Adoption | 12 / 20 |
What we saw
- Starts and answers the MCP handshake: yes (1.7 seconds including download)
- Tools offered: 15
- Last published: 2026-11-05
- Weekly downloads (npm): 1,166
- License: Apache-2.0 · Direct dependencies: 3
Want this server set up for your team?
ConnectATD builds AI automation and wires up MCP servers like this one. Tell us what you are trying to do.
Talk to ConnectATDTools and how they are described
Agents choose tools from these descriptions. Under each name you can see how many of its parameters are explained.
vault_listno parametersList the 1Password vaults accessible to the service account (limited to the allow-listed vaults when OP_MCP_ALLOWED_VAULTS / --allowed-vaults is configured). Returns vault IDs, names, descriptions, and types.
item_lookup3 of 3 parameters describedSearch for items within a 1Password vault by title substring. Returns item IDs, titles, categories, and vault IDs.
item_delete2 of 2 parameters describedPermanently delete an item from a 1Password vault. This action cannot be undone.
item_get4 of 4 parameters describedRetrieve a full 1Password item — title, category, tags, notes, and all fields (id, title, type, section). Secret-bearing field values (passwords and other concealed fields, SSH private keys, one-time-password seeds, card numbers) are hidden unless reveal is true; only known non-secret types (text, URL, email, phone, date, menu, card type, address, reference) are shown by default. Accepts a secret reference (op://vault/item/field) or vault ID + item ID. Revealing a secret puts it in the model context/transcript — to USE a secret in a command or API call, prefer op_run with op:// references instead.
item_edit8 of 8 parameters describedEdit an existing 1Password item. Update the title, notes (pass an empty string to clear notes), tags, website URL, upsert fields, or remove fields. Only referenced fields are changed; all others are preserved.
item_list1 of 1 parameter describedList all items in a 1Password vault, returning id, title, category, tags, and updatedAt for each. Never returns secret values.
item_archive2 of 2 parameters describedArchive an item in a 1Password vault. The item is moved to the archive and hidden from regular views, rather than being permanently deleted.
note_create5 of 5 parameters describedCreate a Secure Note item in a 1Password vault, with optional tags and custom fields.
password_create9 of 9 parameters describedCreate a new password/login item in a 1Password vault with optional username, URL, tags, and notes.
password_read5 of 5 parameters describedRetrieve a secret from 1Password using either a secret reference (op://vault/item/field) or vault ID + item ID. Supports field selection and optional value reveal (defaults to metadata-only). Revealing a secret puts it in the model context/transcript — to USE a secret in a command or API call, prefer op_run with op:// references instead.
password_update5 of 5 parameters describedUpdate (rotate) a password or concealed field on an existing 1Password item. If the target field does not exist, it will be created.
password_generate4 of 4 parameters describedGenerate a cryptographically secure random password with configurable length and character types. Uses rejection sampling for unbiased randomness.
password_generate_memorable5 of 5 parameters describedGenerate a memorable passphrase from random dictionary words with optional number and symbol suffixes. Uses a ~500-word curated list for good entropy.
op_run7 of 7 parameters describedRun a local shell command with 1Password secrets injected as environment variables. Resolved values go only into the child process environment, are never logged by the server, and are redacted from stdout/stderr on a best-effort basis (including common encodings such as JSON, URL and base64) — a command that deliberately transforms or transmits a secret can still expose it. This is the preferred way to USE a secret in a command, API call, or script: prefer op_run with op://vault/item/field references in `env` over reading a secret with password_read/item_get and pasting it into a command yourself, which would put the plaintext in the model context and transcript.
op_check_ref1 of 1 parameter describedValidate an op://vault/item/field secret reference and return only non-secret metadata (vault name, item title, field label, field type) confirming it resolves — the field VALUE is never returned. Use this to check a reference is correct before using it with op_run; do not use password_read/item_get with reveal just to check a reference exists.
Tested in a sandbox with an empty environment: no API keys, a throwaway folder, nothing from a real account. A passing test shows the server runs and responds. It is not a security audit.