Role-Based Access Control in the Archestra Platform

Archestra uses a role-based access control (RBAC) system to manage user permissions across every resource in the platform. Permissions follow a resource:action format — for example, agent:create allows creating agents, mcpGateway:update allows modifying MCP gateway configuration, and llmProxy:read allows viewing LLM proxies. This model gives platform administrators precise control over what each user can see and do without granting blanket elevated access.

​

Predefined Roles

The following roles are built into Archestra and cannot be modified or deleted.

​

Editor Permissions

View full Editor permission matrix

Resource	Actions
Agents	read, create, update, delete, team-admin
Skills	read, create, update, delete, team-admin, execute
Agent Triggers	read, create, update, delete
Scheduled Tasks	read, create, update, delete
LLM Proxies	read, create, update, delete, team-admin
LLM Provider API Keys	read, create, update, delete
LLM Virtual Keys	read, create, update, delete
LLM OAuth Clients	read, create, update, delete
LLM Models	read, update
LLM Limits	read, create, update, delete
Optimization Rules	read, create, update, delete
LLM Costs	read
MCP Gateways	read, create, update, delete, team-admin
Tools & Policies	read, create, update, delete
MCP Registry	read, create, update, delete, team-admin
MCP Server Installations	read, create, update, delete
MCP Server Installation Requests	read, create, update, delete
Environments	admin
GitHub App Configurations	read, create, update, delete
Knowledge Files	read, create, update, delete
Knowledge Sources	read, create, update, delete, query
Chats	read, create, update, delete
Logs	read
API Keys	read, create, delete
LLM Settings	read, update
Knowledge Settings	read, update
Users	read
Invitations	read
Roles	read
Teams	read
Identity Providers	read
Secrets	read
Organization Settings	read, update
Site Notifications	read
Chat Agent Picker	enable
Chat Provider Settings	enable
Chat Expand Tool Calls	enable

​

Member Permissions

View full Member permission matrix

Resource	Actions
Agents	read, create, update, delete
Skills	read, create, update, delete, execute
Scheduled Tasks	read, create, update, delete
LLM Proxies	read, create, update, delete
LLM Provider API Keys	read
LLM Virtual Keys	read
LLM OAuth Clients	read
LLM Models	read
MCP Gateways	read, create, update, delete
Tools & Policies	read
MCP Registry	read
MCP Server Installations	read, create, delete
MCP Server Installation Requests	read, create, update
Knowledge Files	read
Knowledge Sources	read, query
Chats	read, create, update, delete
API Keys	read, create, delete
Teams	read
Site Notifications	read
Simple View	enable
Chat Agent Picker	enable
Chat Provider Settings	enable
Chat Expand Tool Calls	enable

​

Custom Roles

Users with the ac:create permission can create custom roles by selecting specific permission combinations. Custom roles enable fine-grained access control tailored to your organization’s needs. You can only grant permissions that you already possess — this prevents privilege escalation.

​

Available Permissions

Access Control

Permission	Description
ac:read	View custom roles and their permissions
ac:create	Create new custom roles
ac:update	Modify custom role permissions
ac:delete	Delete custom roles

Agents & Skills

Permission	Description
agent:read	View and list agents
agent:create	Create new agents
agent:update	Modify agent configuration and settings
agent:delete	Delete agents
agent:team-admin	Manage team assignments for agents
agent:admin	Full administrative control over all agents, bypassing team restrictions
skill:read	View and use agent skills within your scope
skill:create	Create new agent skills
skill:update	Modify agent skills and their team assignments
skill:delete	Delete agent skills
skill:team-admin	Manage team assignments for agent skills
skill:admin	Full administrative control over all agent skills
skill:execute	Execute skill scripts
agentSettings:read	View agent settings (default model, default agent, security engine, file uploads)
agentSettings:update	Modify agent settings (default model, default agent, security engine, file uploads)
agentTrigger:read	View agent trigger configurations
agentTrigger:create	Set up new agent triggers
agentTrigger:update	Modify agent trigger configurations
agentTrigger:delete	Remove agent triggers
scheduledTask:read	View scheduled tasks and their run history
scheduledTask:create	Create new scheduled tasks and trigger runs
scheduledTask:update	Modify scheduled task configuration
scheduledTask:delete	Delete scheduled tasks
scheduledTask:admin	View and manage all scheduled tasks, not just your own

LLM Proxy

Permission	Description
llmProxy:read	View and list LLM proxies
llmProxy:create	Create new LLM proxies
llmProxy:update	Modify LLM proxy configuration
llmProxy:delete	Delete LLM proxies
llmProxy:team-admin	Manage team assignments for LLM proxies
llmProxy:admin	Full administrative control over all LLM proxies
llmProviderApiKey:read	View LLM provider API keys
llmProviderApiKey:create	Add new LLM provider API keys
llmProviderApiKey:update	Modify LLM provider API key configuration and visibility
llmProviderApiKey:delete	Remove LLM provider API keys
llmProviderApiKey:admin	Manage all LLM provider API keys, including org-wide keys
llmVirtualKey:read	View LLM virtual keys
llmVirtualKey:create	Create LLM virtual keys
llmVirtualKey:update	Modify LLM virtual keys and their visibility
llmVirtualKey:delete	Delete LLM virtual keys
llmVirtualKey:admin	Manage all LLM virtual keys and view every scope
llmOauthClient:read	View LLM OAuth client registrations
llmOauthClient:create	Create LLM OAuth client registrations
llmOauthClient:update	Modify LLM OAuth client registrations
llmOauthClient:delete	Delete LLM OAuth client registrations
llmOauthClient:admin	Manage all LLM OAuth client registrations
llmSettings:read	View LLM settings (compression, cleanup interval)
llmSettings:update	Modify LLM settings
llmModel:read	View synced LLM models and capabilities
llmModel:update	Modify LLM model pricing and modality settings
llmLimit:read	View token usage limits
llmLimit:create	Create new usage limits
llmLimit:update	Modify existing usage limits
llmLimit:delete	Remove usage limits
llmCost:read	View LLM usage cost statistics and analytics
optimizationRule:read	View optimization rules
optimizationRule:create	Create new optimization rules
optimizationRule:update	Modify optimization rules
optimizationRule:delete	Remove optimization rules

MCP Gateways & Registry

Permission	Description
mcpGateway:read	View and list MCP gateways
mcpGateway:create	Create new MCP gateways
mcpGateway:update	Modify MCP gateway configuration
mcpGateway:delete	Delete MCP gateways
mcpGateway:team-admin	Manage team assignments for MCP gateways
mcpGateway:admin	Full administrative control over all MCP gateways
mcpRegistry:read	Browse the MCP server registry
mcpRegistry:create	Add servers to the MCP registry
mcpRegistry:update	Modify MCP registry entries
mcpRegistry:delete	Remove servers from the MCP registry
mcpRegistry:team-admin	Manage team assignments for MCP registry entries
mcpServerInstallation:read	View installed MCP servers and their status
mcpServerInstallation:create	Install MCP servers from the registry
mcpServerInstallation:update	Modify installed MCP server configuration
mcpServerInstallation:delete	Uninstall MCP servers
mcpServerInstallation:admin	Approve or manage all MCP server installations
mcpServerInstallationRequest:read	View MCP server installation requests
mcpServerInstallationRequest:create	Submit requests to install MCP servers
mcpServerInstallationRequest:update	Add notes to installation requests
mcpServerInstallationRequest:delete	Delete installation requests
mcpServerInstallationRequest:admin	Approve or decline installation requests
toolPolicy:read	View tools, tool invocation policies, and trusted data policies
toolPolicy:create	Register tools and create security policies
toolPolicy:update	Modify tools, tool configuration, and security policies
toolPolicy:delete	Remove tools and security policies

Knowledge Base

Permission	Description
knowledgeFile:read	View uploaded Knowledge Files
knowledgeFile:create	Upload Knowledge Files
knowledgeFile:update	Modify Knowledge File visibility and agent access
knowledgeFile:delete	Delete Knowledge Files
knowledgeFile:admin	View all Knowledge Files, bypassing visibility restrictions
knowledgeSource:read	View Knowledge Bases and Connectors
knowledgeSource:create	Create Knowledge Bases and Connectors
knowledgeSource:update	Modify Knowledge Bases and Connectors
knowledgeSource:delete	Delete Knowledge Bases and Connectors
knowledgeSource:query	Query knowledge sources for information retrieval
knowledgeSource:admin	View all Knowledge Bases and Connectors, bypassing visibility restrictions
knowledgeSettings:read	View knowledge settings (embedding and reranking models)
knowledgeSettings:update	Modify knowledge settings (embedding and reranking models)

Users, Teams & Organization

Permission	Description
member:read	View organization members and their roles
member:create	Add new members to the organization
member:update	Change member roles and settings
member:delete	Remove members from the organization
team:read	View teams and their members
team:create	Create new teams
team:update	Modify team settings
team:delete	Delete teams
team:admin	Manage team membership (add/remove members)
invitation:create	Send invitations to new users
invitation:cancel	Cancel pending invitations
organizationSettings:read	View organization settings
organizationSettings:update	Customize organization appearance, authentication, etc.
identityProvider:read	View identity provider configurations (SSO)
identityProvider:create	Set up new identity providers
identityProvider:update	Modify identity provider settings
identityProvider:delete	Remove identity providers
environment:admin	Create, edit, and delete deployment environments (everyone can view them)
environment:deploy-to-restricted	Deploy catalog items to restricted environments
githubAppConfig:read	View GitHub App configurations
githubAppConfig:create	Create GitHub App configurations
githubAppConfig:update	Modify GitHub App configurations
githubAppConfig:delete	Delete GitHub App configurations
auditLog:read	View the organization-wide audit log of administrative actions

API Keys, Secrets & Chat

Permission	Description
apiKey:read	View API keys
apiKey:create	Create API keys
apiKey:delete	Delete API keys
serviceAccount:read	View service accounts
serviceAccount:create	Create service accounts
serviceAccount:update	Modify service accounts
serviceAccount:delete	Delete service accounts
secret:read	View secrets manager configuration
secret:update	Modify secrets manager settings and test connectivity
chat:read	View and access chat conversations
chat:create	Start new chat conversations
chat:update	Edit chat messages and conversation settings
chat:delete	Delete chat conversations
chatAgentPicker:enable	Show agent picker in chat
chatExpandToolCalls:enable	Allow expanding tool call details in chat
chatProviderSettings:enable	Show model and API key selectors in chat
simpleView:enable	Sidebar is collapsed by default on page load
siteNotification:read	View site-wide notifications
siteNotification:create	Create new site notifications
siteNotification:update	Modify site notifications
siteNotification:delete	Delete site notifications
log:read	View LLM proxy and MCP tool call logs

​

Scoped Resources

Some resources use a two-step authorization model. RBAC grants a base action (read, create, update, delete), and then runtime scope rules further restrict which records a user can actually see or modify. The most common scopes are:

• personal — owned by one user
• team — shared with one or more teams
• org — shared across the entire organization

​

Agents, MCP Gateways, and LLM Proxies

These three resources share the same scope model:

​

Visibility-Scoped Credentials

llmProviderApiKey and llmVirtualKey support personal, team, and org scope but use different elevated permissions. Team records require membership in the selected team; organization-wide records require the resource-specific admin permission such as llmProviderApiKey:admin. These resources do not use :team-admin.

​

Chat Access

Chat access requires both chat:read and agent:read because chat is agent-backed — a user must be able to access at least one agent or profile context to start or use chat. The chatAgentPicker:enable, chatProviderSettings:enable, and chatExpandToolCalls:enable permissions are UI toggles only and do not grant access to provider credentials or model catalogs on their own.

​

Team-Based Access Control

Teams allow you to share agents, MCP gateways, LLM proxies, and MCP servers with specific groups of users. Team-based access works alongside RBAC — roles control what actions a user can perform, and teams control which records they can see.

​

Default Team

New users are automatically added to the Default Team when they accept an invitation. This ensures all users have immediate access to any Archestra resources assigned to this team.

​

Team Access Rules

Team-based access also extends to related artifacts such as interaction logs, policies, and tool assignments — members can only view these for agents and MCP servers they have access to.

​

Best Practices

• Principle of Least Privilege — Grant only the minimum permissions necessary for a user’s role.
• Team-based organization — Combine roles with teams for fine-grained resource access rather than granting broad org-wide permissions.
• Regular review — Periodically review custom roles and team memberships to ensure they align with current needs and security requirements.
• Clear naming — Use descriptive names for custom roles that indicate their purpose, such as Agent-Manager, Read-Only-Analyst, or Tool-Developer.