RuntimeCapabilities — @personaai/runtime

Generated field table for RuntimeCapabilities — do not edit

Generated from sdk/runtime/dist/typedoc.json — do not edit.

`RuntimeCapabilities`

Every one of these is false unless explicitly enabled — always-off by default so upgrading this package never silently exposes new surface to whoever resolveUser accepts. These resources are Project-level configuration (LLM provider credentials, skill/knowledge-base/vector-store management, security audit logs, an agent-building co-pilot) rather than things an end user does in a chat session — most hosts should manage them via @personaai/sdk directly from their own admin surface and never turn these on. Turn one on only if you specifically want it reachable through whatever resolveUser gates (which may not be "any logged-in end user" — that's your call).

Source: types/options.ts:36

Interface

Field	Type	Description
`agentsWrite?`	`boolean`	Full Agent CRUD (create/get/update/delete/bulk-delete) beyond the always-on read-only `GET /agents` list. — _ `types/options.ts:38`_
`mcps?`	`boolean`	Full MCP server CRUD + testConnection/readResource/callTool, beyond the always-on `/mcps/:id/oauth/*` routes. — _ `types/options.ts:40`_
`providers?`	`boolean`	LLM provider configuration — holds API keys. — _ `types/options.ts:42`_
`skills?`	`boolean`	Skill authoring. — _ `types/options.ts:44`_
`knowledge?`	`boolean`	Knowledge base CRUD, document upload/search. — _ `types/options.ts:46`_
`stores?`	`boolean`	Vector store CRUD and file read/write. — _ `types/options.ts:48`_
`auditLogs?`	`boolean`	Security/compliance audit log read access. — _ `types/options.ts:50`_
`architect?`	`boolean`	The Architect co-pilot — builds/edits Agents on the caller's behalf via tool calls. — _ `types/options.ts:52`_
`workflowsWrite?`	`boolean`	Full Workflow CRUD, draft authoring, versions, and Mermaid export beyond the always-on execution routes. — _ `types/options.ts:54`_

RuntimeBufferedResponse — @personaai/runtime

RuntimeHooks — @personaai/runtime