epic-permissions
Explains how to implement and validate RBAC permissions for applications built on the Epic Stack.
Install
mkdir -p .claude/skills/epic-permissions && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/1777" && unzip -o skill.zip -d .claude/skills/epic-permissions && rm skill.zipInstalls to .claude/skills/epic-permissions
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
Guide on RBAC system and permissions for Epic StackKey capabilities
- →Implement role-based access control (RBAC)
- →Validate permissions on server-side or client-side
- →Create new permissions or roles
- →Restrict access to routes or actions
- →Implement granular permissions like 'own' vs 'any'
How it works
The system uses an RBAC model where users have roles, and roles have permissions, with a user's permissions being the union of all permissions from their roles.
Inputs & outputs
When to use epic-permissions
- →Implementing RBAC
- →Validating permissions
- →Restricting route access
- →Defining permission strings
About this skill
Epic Stack: Permissions
When to use this skill
Use this skill when you need to:
- Implement role-based access control (RBAC)
- Validate permissions on server-side or client-side
- Create new permissions or roles
- Restrict access to routes or actions
- Implement granular permissions (
ownvsany)
Patterns and conventions
Permissions Philosophy
Following Epic Web principles:
Explicit is better than implicit - Always explicitly check permissions. Don't assume a user has access based on implicit rules or hidden logic. Every permission check should be visible and clear in the code.
Example - Explicit permission checks:
// ✅ Good - Explicit permission check
export async function action({ request }: Route.ActionArgs) {
const userId = await requireUserId(request)
// Explicitly check permission - clear and visible
await requireUserWithPermission(request, 'delete:note:own')
// Permission check is explicit and obvious
await prisma.note.delete({ where: { id: noteId } })
}
// ❌ Avoid - Implicit permission check
export async function action({ request }: Route.ActionArgs) {
const userId = await requireUserId(request)
const note = await prisma.note.findUnique({ where: { id: noteId } })
// Implicit check - not clear what permission is being checked
if (note.ownerId !== userId) {
throw new Response('Forbidden', { status: 403 })
}
// What permission does this represent? Not explicit
}
Example - Explicit permission strings:
// ✅ Good - Explicit permission string
const permission: PermissionString = 'delete:note:own'
// Clear: action (delete), entity (note), access (own)
await requireUserWithPermission(request, permission)
// ❌ Avoid - Implicit or unclear permissions
const canDelete = checkUserCanDelete(user, note)
// What permission is this checking? Not explicit
RBAC Model
Epic Stack uses an RBAC (Role-Based Access Control) model where:
- Users have Roles
- Roles have Permissions
- A user's permissions are the union of all permissions from their roles
Permission Structure
Permissions follow the format: action:entity:access
Components:
action: The allowed action (create,read,update,delete)entity: The entity being acted upon (user,note, etc.)access: The access level (own,any,own,any)
Examples:
create:note:own- Can create own notesread:note:any- Can read any notedelete:user:any- Can delete any user (admin)update:note:own- Can update only own notes
Prisma Schema
Models:
model Permission {
id String @id @default(cuid())
action String // e.g. create, read, update, delete
entity String // e.g. note, user, etc.
access String // e.g. own or any
description String @default("")
roles Role[]
@@unique([action, entity, access])
}
model Role {
id String @id @default(cuid())
name String @unique
description String @default("")
users User[]
permissions Permission[]
}
model User {
id String @id @default(cuid())
// ...
roles Role[]
}
Validate Permissions Server-Side
Require specific permission:
import { requireUserWithPermission } from '#app/utils/permissions.server.ts'
export async function action({ request }: Route.ActionArgs) {
const userId = await requireUserWithPermission(
request,
'delete:note:own', // Throws 403 error if doesn't have permission
)
// User has the permission, continue...
}
Require specific role:
import { requireUserWithRole } from '#app/utils/permissions.server.ts'
export async function loader({ request }: Route.LoaderArgs) {
const userId = await requireUserWithRole(request, 'admin')
// User has admin role, continue...
}
Conditional permissions (own vs any) - explicit:
export async function action({ request }: Route.ActionArgs) {
const userId = await requireUserId(request)
// Explicitly determine ownership
const note = await prisma.note.findUnique({
where: { id: noteId },
select: { ownerId: true },
})
const isOwner = note.ownerId === userId
// Explicitly check the appropriate permission based on ownership
await requireUserWithPermission(
request,
isOwner ? 'delete:note:own' : 'delete:note:any', // Explicit permission string
)
// Permission check is explicit and clear
// Proceed with deletion...
}
Validate Permissions Client-Side
Check if user has permission:
import { userHasPermission, useOptionalUser } from '#app/utils/user.ts'
export default function NoteRoute({ loaderData }: Route.ComponentProps) {
const user = useOptionalUser()
const isOwner = user?.id === loaderData.note.ownerId
const canDelete = userHasPermission(
user,
isOwner ? 'delete:note:own' : 'delete:note:any',
)
return (
<div>
{canDelete && (
<button onClick={handleDelete}>Delete</button>
)}
</div>
)
}
Check if user has role:
import { userHasRole } from '#app/utils/user.ts'
export default function AdminRoute() {
const user = useOptionalUser()
const isAdmin = userHasRole(user, 'admin')
if (!isAdmin) {
return <div>Access Denied</div>
}
return <div>Admin Panel</div>
}
Create New Permissions
En Prisma Studio o seed:
// prisma/seed.ts
await prisma.permission.create({
data: {
action: 'create',
entity: 'post',
access: 'own',
description: 'Can create their own posts',
roles: {
connect: { name: 'user' },
},
},
})
Permiso con múltiples niveles de acceso:
await prisma.permission.createMany({
data: [
{
action: 'read',
entity: 'post',
access: 'own',
description: 'Can read own posts',
},
{
action: 'read',
entity: 'post',
access: 'any',
description: 'Can read any post',
},
],
})
Assign Roles to Users
When creating user:
const user = await prisma.user.create({
data: {
email,
username,
roles: {
connect: { name: 'user' }, // Assign 'user' role
},
},
})
Assign multiple roles:
await prisma.user.update({
where: { id: userId },
data: {
roles: {
connect: [{ name: 'user' }, { name: 'moderator' }],
},
},
})
Permissions and Roles Seed
Seed example:
// prisma/seed.ts
// Create permissions
const permissions = await Promise.all([
// User permissions
prisma.permission.create({
data: {
action: 'create',
entity: 'note',
access: 'own',
description: 'Can create own notes',
},
}),
prisma.permission.create({
data: {
action: 'read',
entity: 'note',
access: 'own',
description: 'Can read own notes',
},
}),
prisma.permission.create({
data: {
action: 'update',
entity: 'note',
access: 'own',
description: 'Can update own notes',
},
}),
prisma.permission.create({
data: {
action: 'delete',
entity: 'note',
access: 'own',
description: 'Can delete own notes',
},
}),
// Admin permissions
prisma.permission.create({
data: {
action: 'delete',
entity: 'user',
access: 'any',
description: 'Can delete any user',
},
}),
])
// Create roles
const userRole = await prisma.role.create({
data: {
name: 'user',
description: 'Standard user',
permissions: {
connect: permissions.slice(0, 4).map((p) => ({ id: p.id })),
},
},
})
const adminRole = await prisma.role.create({
data: {
name: 'admin',
description: 'Administrator',
permissions: {
connect: permissions.map((p) => ({ id: p.id })),
},
},
})
Permission Type
Type-safe permission strings:
import { type PermissionString } from '#app/utils/user.ts'
// Tipo: 'create:note:own' | 'read:note:own' | etc.
const permission: PermissionString = 'delete:note:own'
Parsear permission string:
import { parsePermissionString } from '#app/utils/user.ts'
const { action, entity, access } = parsePermissionString('delete:note:own')
// action: 'delete'
// entity: 'note'
// access: ['own']
Common examples
Example 1: Proteger action con permiso
// app/routes/users/$username/notes/$noteId.tsx
export async function action({ request }: Route.ActionArgs) {
const userId = await requireUserId(request)
const formData = await request.formData()
const { noteId } = Object.fromEntries(formData)
const note = await prisma.note.findFirst({
select: { id: true, ownerId: true, owner: { select: { username: true } } },
where: { id: noteId },
})
if (!note) {
throw new Response('Not found', { status: 404 })
}
const isOwner = note.ownerId === userId
// Validate permiso según si es propietario o no
await requireUserWithPermission(
request,
isOwner ? 'delete:note:own' : 'delete:note:any',
)
await prisma.note.delete({ where: { id: note.id } })
return redirect(`/users/${note.owner.username}/notes`)
}
Example 2: Mostrar UI condicional basada en permisos
export default function NoteRoute({ loaderData }: Route.ComponentProps) {
const user = useOptionalUser()
const isOwner = user?.id === loaderData.note.ownerId
const canDelete = userHasPermission(
user,
isOwner ? 'delete:note:own' : 'delete:note:any',
)
const canEdit = userHasPermission(
user,
isOwner ? 'update:note:own' : 'update:note:any',
)
return (
<div>
<h1>{loaderData.note.title}</h1>
<p>{loaderData.note.content}</p>
{(canEdit || canDelete) && (
<div className="flex gap-2">
{canEdit && (
<Link to="edit">
<Button>Edit</Button>
</Link>
)}
{canDelete && (
<DeleteNoteButton noteId={loaderData.note.id} />
)}
</div>
)}
</div>
)
}
Example 3: Ruta solo para admin
// app/routes/admin/users.tsx
export async function loader({ request }: Route.LoaderArgs) {
await requireUserWithRole(request, 'admin')
const users = await pr
---
*Content truncated.*
Limitations
- →Permissions must be explicitly checked
- →Permissions must be validated on the server-side
- →The 'own' vs 'any' access level must be explicitly verified
How it compares
This skill emphasizes explicit permission checks and structured permission strings, contrasting with implicit or unclear access control logic.
Compared to similar skills
epic-permissions side by side with the closest alternatives in the catalog.
| Skill | Installs | Updated | Safety | Difficulty |
|---|---|---|---|---|
| epic-permissions (this skill) | 3 | 6mo | No flags | Intermediate |
| epic-auth | 1 | 6mo | Review | Advanced |
| linear-security-basics | 1 | 27d | Review | Intermediate |
| middleware-protection | 1 | 6mo | Review | Intermediate |
Try saying
Example prompts that trigger this skill in your AI assistant.
More by epicweb-dev
View all by epicweb-dev →You might also like
epic-auth
epicweb-dev
Guide on authentication, sessions, OAuth, 2FA, and passkeys for Epic Stack
linear-security-basics
jeremylongshore
Secure API key management and OAuth best practices for Linear. Use when setting up authentication securely, implementing OAuth flows, or hardening Linear integrations. Trigger with phrases like "linear security", "linear API key security", "linear OAuth", "secure linear integration", "linear secrets management".
middleware-protection
dadbodgeoff
Protect routes with Next.js middleware. Check authentication once, protect routes declaratively. Supports public routes, protected routes, and role-based access.
auth-patterns
davepoon
This skill should be used when the user asks about "authentication in Next.js", "NextAuth", "Auth.js", "middleware auth", "protected routes", "session management", "JWT", "login flow", or needs guidance on implementing authentication and authorization in Next.js applications.
jwt-auth
dadbodgeoff
Implement secure JWT authentication with refresh token rotation, secure storage, and automatic renewal. Use when building authentication for SPAs, mobile apps, or APIs that need stateless auth with refresh capabilities.
fix-dependabot-alerts
microsoft
Fix Dependabot security alerts by updating vulnerable npm dependencies. Use when the user mentions "dependabot", "security alerts", "vulnerability", "CVE", or wants to update packages with security issues.