Skip to main content
Glama
wundervault

wundervault

Official

vault_new_key

Create a new service API key using an existing vaulted credential, then store the new key in the owner’s vault for passkey approval so agents never see it.

Instructions

Make a NEW key at a service (for example a new API key at Resend or Cloudflare), using a vaulted key that is allowed to create keys there. The helper sends your request, takes the new key out of the answer and puts it straight into the owner's vault, where it waits for the owner. You never see the new key: you get its id and name at the service. No agent can use it until the owner saves it with their passkey and sets it up, so tell the owner it is waiting under Needs you.

Works only at these exact addresses (POST): Resend https://api.resend.com/api-keys; Cloudflare https://api.cloudflare.com/client/v4/user/tokens and /client/v4/accounts/{account_id}/tokens; SendGrid https://api.sendgrid.com/v3/api_keys; OpenAI https://api.openai.com/v1/organization/admin_api_keys and /v1/organization/projects/{project_id}/service_accounts; Twilio https://api.twilio.com/2010-04-01/Accounts/{account_sid}/Keys.json; OpenRouter https://openrouter.ai/api/v1/keys. vault_call refuses these addresses. Anywhere else, the owner makes the key and adds it to the vault by hand.

ROTATING a key: pass replaces_entry_id (the secret the new key replaces). The order is: (0) list the keys at the service and note the id of the OLD key, the one in use now; (1) vault_new_key, giving the new key the same permissions; (2) tell the owner and wait until they press Replace; (3) make a harmless call with that secret to prove the new value works; (4) only then delete the OLD key at the service with vault_call (DELETE, by the OLD id you noted, never the id vault_new_key returned); (5) tell the owner. Never delete the old key before the new one has worked. Rolling a token in place (Cloudflare PUT .../tokens/{token_id}/value) is refused: it ends the old value at once.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "entry_id": "<entry_id>",
      "request": "<request>",
      "save_as": "<save_as>",
      "purpose": "<purpose>"
    },
    "name": "vault_new_key"
  }
}

Input Schema

NameRequiredDescriptionDefault
purposeYesWhy this is needed (audit log, and shown to the owner).
requestYesAs for vault_call: {origin:{scheme:"https",host,port:443}, method:"POST", path, body?, headers?}. The body is what the service wants for a new key (its name, permissions).
save_asYesName for the new secret in the vault, 1 to 80 characters.
entry_idYesVault entry holding the key that is allowed to create keys at the service.
replaces_entry_idNoOptional. The entry id of the secret this new key replaces (one you have). The owner can then replace its value in one step; its name, rules and agents stay as they are.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Addedv2.5.0

TDQS

A4.5/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries the full burden and does it well: it discloses the security model (the new key goes straight to the owner's vault, no agent can use it until the owner saves it with a passkey), what the caller receives (id and name, not the secret), and that Cloudflare in-place token rotation is refused. It does not cover failure or retry semantics — e.g. whether a retried call creates a duplicate key at the service — which is the main remaining gap for a creation tool.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Three clearly partitioned paragraphs — purpose and return behavior, allowed addresses, rotation procedure — with the core purpose front-loaded. Dense but essentially waste-free; the rotation block is long, though every step is load-bearing for the coordination with the owner.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a five-parameter nested-object tool with no annotations and no output schema, the description supplies nearly everything an agent needs: scope, endpoints, what is returned, and the multi-party workflow. Error handling and idempotency on partial failure are the only notable omissions.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the baseline would be 3, but the description adds real meaning beyond the schema, particularly for replaces_entry_id: it explains that passing it triggers a one-step Replace flow for the owner and that the new key should be given the same permissions. It also clarifies that request.body is the service's own payload for a new key.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description opens with a specific verb and resource — make a NEW key at a service — and immediately scopes it to producing a credential the agent never sees. It explicitly differentiates itself from the sibling vault_call by noting that vault_call refuses these same addresses.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

It enumerates the exact allowed POST endpoints per provider and states the negative case: anywhere else, the owner makes the key by hand. It also spells out a five-step rotation procedure for replaces_entry_id, including the ordering constraint that the old key must not be deleted until the new one has been proven.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.