Skip to main content

Vault Adapter

The Vault adapter provides simplified file access for agents and users by wrapping DLFS with a fixed drive name. Instead of specifying a drive on every call, vault operations always target the "health-vault" drive.

Why Use Vault?

When agents need file access, giving them the full DLFS API means they must know which drive to use. The Vault adapter removes that decision — agents get simple read, write, list, mkdir, and delete tools without needing to manage drive names.

This is especially useful for:

  • Health data agents that store and retrieve personal health documents
  • Any domain where agents should work in a single, well-known file space

Operations

All operations work on the "health-vault" drive. Only a path is needed.

Read

{ "operation": "vault:read", "input": { "path": "documents/referral.json" } }

Returns {content, encoding, size} — identical to dlfs:read.

Write

// Inline content
{ "operation": "vault:write", "input": { "path": "profile.json", "content": "{...}" } }

// From asset
{ "operation": "vault:write", "input": { "path": "scan.pdf", "asset": "/a/0x1234..." } }

List

{ "operation": "vault:list", "input": { "path": "documents" } }

Omit path to list the vault root.

Mkdir

{ "operation": "vault:mkdir", "input": { "path": "documents/lab-results" } }

Delete

{ "operation": "vault:delete", "input": { "path": "documents/temp.txt" } }

Operations Reference

OperationInputDescription
vault:readpathRead file content
vault:writepath, content? or asset?Write file content
vault:listpath?List directory entries
vault:mkdirpathCreate a directory
vault:deletepathDelete a file or empty directory

Giving Agents Vault Access

Add vault operations to an agent's tool list:

{
"operation": "agent:create",
"input": {
"agentId": "health-assistant",
"config": {
"systemPrompt": "You are a health data assistant...",
"tools": [
"v/ops/vault/read",
"v/ops/vault/write",
"v/ops/vault/list",
"v/ops/vault/mkdir"
],
"defaultTools": false
}
}
}

Example: Agent Processing Health Data

1. Agent calls vault:list → discovers available health documents
2. Agent calls vault:read path="lab-results/blood-panel.json" → retrieves data
3. Agent processes content, produces analysis
4. Agent calls vault:write path="analysis/blood-panel-summary.json" → stores result

All operations are traced to the agent's identity and signed by the user's DLFS key.

  • DLFS — full decentralised file system with multi-drive support and WebDAV
  • Agents — creating and configuring agents with tools