Token导航 LogoToken导航TokenDH.com
开发需要联网github未标认证来源可访问许可证需确认审计异常

bknd-assign-permissionsbknd 分配权限

Agent Skill

bknd-assign-permissions 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

306

周安装

13

GitHub Stars

3

下载量

107
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:bknd-assign-permissions(bknd 分配权限)
来源仓库:https://github.com/cameronapak/bknd-skills
仓库路径:skills/bknd-assign-permissions
安装命令:
npx skills add https://github.com/cameronapak/bknd-skills --skill bknd-assign-permissions
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/cameronapak/bknd-skills --skill bknd-assign-permissions

简介

bknd-assign-permissions 用于配置角色细粒度权限,适合在 Codex、Claude、Cursor、Gemini CLI 中实现 RBAC 权限管理体系。

  • 适用场景包括定义只读/读写权限、设置条件策略及分配 effect 操作类型(allow/deny)。
  • 核心能力是支持简单字符串权限和扩展格式两种表达方式,满足不同安全等级需求。
  • 使用方式必须启用 guard 模块并在 code 模式下操作,UI 仅提供只读查看功能。
  • 创建角色前需确保 auth 和 guard 均已开启,否则无法生效。

SKILL.md

Assign Permissions

Configure detailed permissions for roles using simple strings, extended format with effects, and conditional policies.

Prerequisites

  • Bknd project with code-first configuration
  • Auth enabled (auth: {enabled: true})
  • Guard enabled (guard: {enabled: true})
  • At least one role defined (see bknd-create-role)

When to Use UI Mode

  • Viewing current role permissions
  • Quick permission checks

UI steps: Admin Panel > Auth > Roles > Select role

Note: Permission assignment requires code mode. UI is read-only.

When to Use Code Mode

  • Assigning permissions to roles
  • Adding permission effects (allow/deny)
  • Creating conditional policies
  • Entity-specific permission rules

Code Approach

Step 1: Simple Permission Strings

Assign basic permissions as string array:

import { serve } from "bknd/adapter/bun";
import { em, entity, text } from "bknd";

const schema = em({
  posts: entity("posts", { title: text().required() }),
});

serve({
  connection: { url: "file:data.db" },
  config: {
    data: schema.toJSON(),
    auth: {
      enabled: true,
      guard: { enabled: true },
      roles: {
        editor: {
          implicit_allow: false,
          permissions: [
            "data.entity.read",    // Read any entity
            "data.entity.create",  // Create in any entity
            "data.entity.update",  // Update any entity
            // No delete permission
          ],
        },
      },
    },
  },
});

Available Permissions

PermissionFilterableDescription
data.entity.readYesRead entity records
data.entity.createYesCreate new records
data.entity.updateYesUpdate existing records
data.entity.deleteYesDelete records
data.database.syncNoSync database schema
data.raw.queryNoExecute raw SELECT
data.raw.mutateNoExecute raw INSERT/UPDATE/DELETE

Filterable means you can add conditions/filters via policies.

Step 2: Extended Permission Format

Use objects for explicit allow/deny effects:

{
  roles: {
    moderator: {
      implicit_allow: false,
      permissions: [
        { permission: "data.entity.read", effect: "allow" },
        { permission: "data.entity.update", effect: "allow" },
        { permission: "data.entity.delete", effect: "deny" },  // Explicit deny
      ],
    },
  },
}

Permission Effects

EffectDescription
allowGrant the permission (default)
denyExplicitly block the permission

Deny overrides allow - useful when implicit_allow: true but you want to block specific actions.

Step 3: Conditional Policies

Add policies for fine-grained control:

{
  roles: {
    content_editor: {
      implicit_allow: false,
      permissions: [
        {
          permission: "data.entity.read",
          effect: "allow",
          policies: [
            {
              description: "Only read posts and comments",
              condition: { entity: { $in: ["posts", "comments"] } },
              effect: "allow",
            },
          ],
        },
        {
          permission: "data.entity.create",
          effect: "allow",
          policies: [
            {
              condition: { entity: { $in: ["posts", "comments"] } },
              effect: "allow",
            },
          ],
        },
      ],
    },
  },
}

Policy Structure

{
  description?: string,      // Human-readable (optional)
  condition?: ObjectQuery,   // When policy applies
  effect: "allow" | "deny" | "filter",
  filter?: ObjectQuery,      // Row filter (for effect: "filter")
}

Policy Effects

EffectPurpose
allowGrant when condition met
denyBlock when condition met
filterApply row-level filter to results

Condition Operators

OperatorDescriptionExample
$eqEqual{entity: {$eq: "posts"}}
$neNot equal{entity: {$ne: "users"}}
$inIn array{entity: {$in: ["posts", "comments"]}}
$ninNot in array{entity: {$nin: ["users", "secrets"]}}
$gtGreater than{age: {$gt: 18}}
$gteGreater or equal{level: {$gte: 5}}
$ltLess than{count: {$lt: 100}}
$lteLess or equal{priority: {$lte: 3}}

Step 4: Variable Placeholders

Reference runtime context with @variable:

PlaceholderDescription
@user.idCurrent user's ID
@user.emailCurrent user's email
@user.roleCurrent user's role
@entityCurrent entity name
@idCurrent record ID

Example - user can only update their own profile:

{
  permissions: [
    {
      permission: "data.entity.update",
      effect: "allow",
      policies: [
        {
          condition: { entity: "users", "@id": "@user.id" },
          effect: "allow",
        },
      ],
    },
  ],
}

Step 5: Entity-Specific Permissions

Grant different permissions per entity:

{
  roles: {
    blog_author: {
      implicit_allow: false,
      permissions: [
        // Full CRUD on posts
        {
          permission: "data.entity.read",
          effect: "allow",
          policies: [{ condition: { entity: "posts" }, effect: "allow" }],
        },
        {
          permission: "data.entity.create",
          effect: "allow",
          policies: [{ condition: { entity: "posts" }, effect: "allow" }],
        },
        {
          permission: "data.entity.update",
          effect: "allow",
          policies: [{ condition: { entity: "posts" }, effect: "allow" }],
        },
        {
          permission: "data.entity.delete",
          effect: "allow",
          policies: [{ condition: { entity: "posts" }, effect: "allow" }],
        },

        // Read-only on categories
        {
          permission: "data.entity.read",
          effect: "allow",
          policies: [{ condition: { entity: "categories" }, effect: "allow" }],
        },
      ],
    },
  },
}

Common Patterns

Read-Only Role

{
  roles: {
    viewer: {
      implicit_allow: false,
      permissions: ["data.entity.read"],
    },
  },
}

CRUD Without Delete

{
  roles: {
    contributor: {
      implicit_allow: false,
      permissions: [
        "data.entity.read",
        "data.entity.create",
        "data.entity.update",
        { permission: "data.entity.delete", effect: "deny" },
      ],
    },
  },
}

Admin with Restricted Raw Access

{
  roles: {
    admin: {
      implicit_allow: true,  // Allow all by default
      permissions: [
        // But deny raw database access
        { permission: "data.raw.query", effect: "deny" },
        { permission: "data.raw.mutate", effect: "deny" },
      ],
    },
  },
}

Multi-Entity Role

{
  roles: {
    content_manager: {
      implicit_allow: false,
      permissions: [
        // Content entities: full CRUD
        {
          permission: "data.entity.read",
          effect: "allow",
          policies: [{
            condition: { entity: { $in: ["posts", "pages", "comments", "media"] } },
            effect: "allow",
          }],
        },
        {
          permission: "data.entity.create",
          effect: "allow",
          policies: [{
            condition: { entity: { $in: ["posts", "pages", "comments", "media"] } },
            effect: "allow",
          }],
        },
        {
          permission: "data.entity.update",
          effect: "allow",
          policies: [{
            condition: { entity: { $in: ["posts", "pages", "comments", "media"] } },
            effect: "allow",
          }],
        },
        {
          permission: "data.entity.delete",
          effect: "allow",
          policies: [{
            condition: { entity: { $in: ["posts", "pages", "comments"] } },  // No media delete
            effect: "allow",
          }],
        },
      ],
    },
  },
}

Deny Specific Entity

{
  roles: {
    user: {
      implicit_allow: false,
      permissions: [
        // Can read most entities
        "data.entity.read",
        // But never access secrets entity
        {
          permission: "data.entity.read",
          effect: "deny",
          policies: [{
            condition: { entity: "secrets" },
            effect: "deny",
          }],
        },
      ],
    },
  },
}

Create Helper Function

For complex role definitions:

// helpers/permissions.ts
type EntityPermission = "read" | "create" | "update" | "delete";

function entityPermissions(
  entities: string[],
  actions: EntityPermission[]
) {
  const permMap: Record<EntityPermission, string> = {
    read: "data.entity.read",
    create: "data.entity.create",
    update: "data.entity.update",
    delete: "data.entity.delete",
  };

  return actions.map((action) => ({
    permission: permMap[action],
    effect: "allow" as const,
    policies: [{
      condition: { entity: { $in: entities } },
      effect: "allow" as const,
    }],
  }));
}

// Usage
{
  roles: {
    blog_author: {
      implicit_allow: false,
      permissions: [
        ...entityPermissions(["posts", "comments"], ["read", "create", "update"]),
        ...entityPermissions(["categories", "tags"], ["read"]),
      ],
    },
  },
}

Verification

Test permission assignments:

1. Login as user with role:

curl -X POST http://localhost:7654/api/auth/password/login \
  -H "Content-Type: application/json" \
  -d '{"email": "editor@example.com", "password": "password123"}'

2. Test allowed permission:

curl http://localhost:7654/api/data/posts \
  -H "Authorization: Bearer <token>"
# Should return 200 with data

3. Test denied permission:

curl -X DELETE http://localhost:7654/api/data/posts/1 \
  -H "Authorization: Bearer <token>"
# Should return 403 Forbidden

4. Test entity-specific permission:

# If only posts/comments allowed:
curl http://localhost:7654/api/data/users \
  -H "Authorization: Bearer <token>"
# Should return 403 if users entity not in allowed list

Common Pitfalls

Permission Not Taking Effect

Problem: Changed permissions but old behavior persists

Fix: Restart server - role config is loaded at startup:

# Stop and restart
bknd run

Deny Not Overriding

Problem: Deny effect not blocking access

Fix: Check policy condition - deny only applies when condition matches:

// WRONG - no condition, may not match
{ permission: "data.entity.delete", effect: "deny" }

// CORRECT - simple deny at permission level
{
  permissions: [
    "data.entity.read",
    "data.entity.create",
    // Don't include delete at all
  ],
}

Entity Condition Not Matching

Problem: Entity-specific permission not working

Fix: Verify entity name matches exactly:

// WRONG - entity name case matters
{ condition: { entity: "Posts" } }

// CORRECT - use exact entity name
{ condition: { entity: "posts" } }

Multiple Policies Conflict

Problem: Confusing behavior with multiple policies

Fix: Understand evaluation order - first matching policy wins:

{
  policies: [
    // More specific first
    { condition: { entity: "secrets" }, effect: "deny" },
    // General fallback last
    { effect: "allow" },
  ],
}

Variable Placeholder Not Resolving

Problem: @user.id appearing literally in filter

Fix: Variables only work in filter and condition fields:

// CORRECT usage
{
  condition: { "@id": "@user.id" },  // Works
  filter: { user_id: "@user.id" },   // Works
}

DOs and DON'Ts

DO:

  • Start with minimal permissions, add as needed
  • Use $in operator for multiple entities
  • Test each permission after adding
  • Use descriptive policy descriptions
  • Prefer explicit permissions over implicit_allow

DON'T:

  • Grant data.raw.* to non-admin roles (SQL injection risk)
  • Use implicit_allow: true with deny policies (confusing)
  • Forget to restart server after config changes
  • Mix simple strings and extended format unnecessarily
  • Over-complicate with too many nested policies

Related Skills

  • bknd-create-role - Define new roles
  • bknd-row-level-security - Filter data by user ownership
  • bknd-protect-endpoint - Secure specific endpoints
  • bknd-public-vs-auth - Configure public vs authenticated access
  • bknd-setup-auth - Initialize authentication system

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

38.57%
按下载量换算41

Claude

28.76%
按下载量换算31

Cursor

19.36%
按下载量换算21

Gemini CLI

9.09%
按下载量换算10

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

未通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills