Token导航 LogoToken导航TokenDH.com
研究检索需要联网github未标认证来源可访问clear审计通过

security-patterns安全模式

Agent Skill

用于辅助安全审计、权限检查、凭据风险、认证流程和常见漏洞排查。它适合让 Agent 梳理敏感配置、检查依赖风险、分析鉴权逻辑或生成安全复核清单。使用时不能把工具输出直接当最终结论,涉及密钥、令牌、用户数据或生产系统时,应先确认最小权限、脱敏方式和操作边界。

总安装

364

周安装

15

GitHub Stars

21

下载量

119
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

3

许可证

MIT

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:security-patterns(安全模式)
来源仓库:https://github.com/thapaliyabikendra/ai-artifacts
仓库路径:skills/security-patterns
安装命令:
npx skills add https://github.com/thapaliyabikendra/ai-artifacts --skill security-patterns
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。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/thapaliyabikendra/ai-artifacts --skill security-patterns

简介

提供结构化安全检查表与反模式速查表,覆盖身份验证、授权与输入验证等核心领域。

  • 适用于代码评审辅助、架构规范对齐或新人培训材料参考。
  • 包含 FluentValidation、OAuth 2.0 等具体技术实现指引。
  • 需结合项目实际技术栈选择性采纳,通用模板可能存在适配偏差。
  • security-patterns 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Security Patterns for Web Applications

Security patterns and practices for building secure ABP Framework applications.

When to Use

  • Conducting security audits
  • Implementing authentication/authorization
  • Creating threat models (STRIDE)
  • Reviewing code for OWASP Top 10 vulnerabilities
  • Designing permission systems
  • Validating input and sanitizing output

STRIDE Threat Model

Framework

CategoryThreatQuestionMitigation
SpoofingIdentity theftCan attacker impersonate user?Authentication, tokens
TamperingData modificationCan attacker modify data?Integrity checks, signing
RepudiationDenial of actionsCan user deny their actions?Audit logging
Information DisclosureData exposureCan attacker access sensitive data?Encryption, access control
Denial of ServiceAvailability attackCan attacker disrupt service?Rate limiting, scaling
Elevation of PrivilegeUnauthorized accessCan attacker gain higher privileges?Authorization, least privilege

Threat Model Template

## Threat Model: [Feature Name]

**Date**: YYYY-MM-DD
**Reviewer**: [Name]

### Assets
| Asset | Sensitivity | Description |
|-------|-------------|-------------|
| Patient Data | HIGH | PII including medical records |
| User Credentials | CRITICAL | Passwords, tokens |
| Appointment Data | MEDIUM | Scheduling information |

### Threat Analysis
| ID | Category | Threat | Likelihood | Impact | Risk | Mitigation |
|----|----------|--------|------------|--------|------|------------|
| T1 | Spoofing | Attacker impersonates patient | Medium | High | HIGH | OAuth 2.0, MFA |
| T2 | Tampering | Attacker modifies appointment | Low | Medium | LOW | Authorization checks |
| T3 | Info Disclosure | Unauthorized patient data access | Medium | Critical | CRITICAL | Row-level security |
| T4 | Elevation | Receptionist gains admin access | Low | Critical | HIGH | Permission validation |

### Mitigations
| ID | Threat | Control | Status |
|----|--------|---------|--------|
| M1 | T1 | Implement OAuth 2.0 with OpenIddict | Implemented |
| M2 | T3 | Add row-level authorization in AppService | Pending |

OWASP Top 10 Checklist

1. Injection (A01)

// BAD: SQL Injection
var query = $"SELECT * FROM Users WHERE Email = '{email}'";

// GOOD: Parameterized query (EF Core does this automatically)
var user = await _dbContext.Users
    .FirstOrDefaultAsync(u => u.Email == email);

// BAD: Command injection
Process.Start("cmd", $"/c dir {userInput}");

// GOOD: Validate and sanitize input
if (!IsValidPath(userInput))
    throw new BusinessException("Invalid path");

2. Broken Authentication (A02)

// Checklist:
// [ ] Use OAuth 2.0 / OpenIddict
// [ ] Implement token expiry (short-lived access, long-lived refresh)
// [ ] Hash passwords with modern algorithm (BCrypt, Argon2)
// [ ] Implement account lockout after failed attempts
// [ ] Use secure session management
// [ ] Implement MFA for sensitive operations

3. Sensitive Data Exposure (A03)

// BAD: Logging PII
_logger.LogInformation("User {Email} logged in", user.Email);

// GOOD: Log identifiers only
_logger.LogInformation("User {UserId} logged in", user.Id);

// BAD: Returning sensitive data
return new UserDto { PasswordHash = user.PasswordHash };

// GOOD: Exclude sensitive fields
return new UserDto { Id = user.Id, Name = user.Name };

4. Security Misconfiguration (A05)

// Checklist:
// [ ] Disable debug mode in production
// [ ] Remove default credentials
// [ ] Configure CORS properly
// [ ] Set secure headers (CSP, X-Frame-Options)
// [ ] Disable directory listing
// [ ] Keep frameworks updated

5. Broken Access Control (A01)

// BAD: No authorization
public async Task<PatientDto> GetPatientAsync(Guid id)
{
    return await _repository.GetAsync(id);
}

// GOOD: Authorization enforced
[Authorize(ClinicPermissions.Patients.Default)]
public async Task<PatientDto> GetPatientAsync(Guid id)
{
    var patient = await _repository.GetAsync(id);

    // Additional check: Can user access this specific patient?
    await AuthorizationService.CheckAsync(patient, CommonOperations.Get);

    return ObjectMapper.Map<Patient, PatientDto>(patient);
}

ABP Authorization Patterns

Permission Definition

public static class {ProjectName}Permissions
{
    public const string GroupName = "{ProjectName}";

    public static class {Feature}
    {
        public const string Default = GroupName + ".{Feature}";
        public const string Create = Default + ".Create";
        public const string Edit = Default + ".Edit";
        public const string Delete = Default + ".Delete";
        public const string ViewAll = Default + ".ViewAll";
    }
}

AppService Authorization

[Authorize({ProjectName}Permissions.{Feature}.Default)]
public class {Entity}AppService : ApplicationService
{
    [Authorize({ProjectName}Permissions.{Feature}.Create)]
    public async Task<{Entity}Dto> CreateAsync(CreateUpdate{Entity}Dto input)
    {
        // Create logic
    }

    [Authorize({ProjectName}Permissions.{Feature}.Edit)]
    public async Task<{Entity}Dto> UpdateAsync(Guid id, CreateUpdate{Entity}Dto input)
    {
        // Update logic
    }

    [Authorize({ProjectName}Permissions.{Feature}.Delete)]
    public async Task DeleteAsync(Guid id)
    {
        // Delete logic
    }
}

Resource-Based Authorization

public async Task<PatientDto> GetAsync(Guid id)
{
    var patient = await _repository.GetAsync(id);

    // Check if current user can access this specific patient
    if (patient.AssignedDoctorId != CurrentUser.Id)
    {
        await AuthorizationService.CheckAsync(
            {ProjectName}Permissions.{Feature}.ViewAll);
    }

    return ObjectMapper.Map<Patient, PatientDto>(patient);
}

Security Audit Report Template

## Security Audit Report

**Application**: [Name]
**Date**: YYYY-MM-DD
**Auditor**: [Name]
**Risk Level**: Critical | High | Medium | Low

### Executive Summary
[1-2 paragraph overview of findings]

### Findings

#### [VULN-001] [Title]
- **Severity**: Critical | High | Medium | Low
- **Category**: OWASP A01-A10 / STRIDE
- **Location**: `path/to/file.cs:line`
- **Description**: [What the vulnerability is]
- **Impact**: [What could happen if exploited]
- **Reproduction Steps**:
  1. [Step 1]
  2. [Step 2]
- **Recommendation**: [How to fix]
- **Code Example**:

// Vulnerable code [code here]

// Fixed code [code here]


### Summary

| Severity | Count | Fixed | Pending |
| --- | --- | --- | --- |
| Critical | 0 | 0 | 0 |
| High | 0 | 0 | 0 |
| Medium | 0 | 0 | 0 |
| Low | 0 | 0 | 0 |

### Recommendations

1. [Priority recommendation]
2. [Secondary recommendation]

Security Checklist

Authentication

  • [ ] OAuth 2.0 / OpenID Connect implemented
  • [ ] Token expiry configured (access: 15-60 min, refresh: 7-30 days)
  • [ ] Password policy enforced (min length, complexity)
  • [ ] Account lockout after failed attempts
  • [ ] MFA available for sensitive operations
  • [ ] Secure password reset flow

Authorization

  • [ ] All endpoints have [Authorize] attribute
  • [ ] Permissions defined for all operations
  • [ ] Role-based access enforced
  • [ ] Resource-based authorization where needed
  • [ ] No permission bypass vulnerabilities
  • [ ] Least privilege principle applied

Input Validation

  • [ ] All DTOs have FluentValidation
  • [ ] SQL uses parameterized queries (EF Core)
  • [ ] File uploads restricted by type and size
  • [ ] API rate limiting configured
  • [ ] XSS prevention (output encoding)
  • [ ] CSRF protection enabled

Data Protection

  • [ ] PII not logged
  • [ ] Sensitive data encrypted at rest
  • [ ] TLS enforced (HTTPS only)
  • [ ] Secure headers configured
  • [ ] Error messages don't expose internals
  • [ ] Connection strings secured

Audit & Monitoring

  • [ ] Security events logged
  • [ ] Failed auth attempts tracked
  • [ ] Admin actions audited
  • [ ] Anomaly detection configured
  • [ ] Log integrity protected

Authorization Anti-Patterns (Quick Scan)

Use this table for rapid code review scanning:

PatternRisk LevelFix
No [Authorize] on public method🔴 CRITICALAdd [Authorize(Permission)]
[Authorize] only at class level🟡 MEDIUMAdd method-level permissions for mutations
No permission check for bulk operations🔴 HIGHCheck permission per operation or batch
Missing [RequiresTenant] on tenant-specific ops🔴 HIGHAdd [RequiresTenant] attribute
_dataFilter.Disable<IMultiTenant>() without comment🔴 CRITICALAdd justification comment or remove
Hardcoded secrets in code🔴 CRITICALUse configuration/secrets management
PII in log messages🟡 MEDIUMLog identifiers only, not PII

Multi-Tenancy Security

Dangerous Pattern: Disabling Tenant Filter

// ⚠️ DANGEROUS: Cross-tenant data exposure risk!
using (_dataFilter.Disable<IMultiTenant>())
{
    // This query now sees ALL tenants' data!
    var exists = await _repository.AnyAsync(x => x.Code == code);
}

Risks:

  • Cross-tenant data leakage
  • Incorrect validation results (e.g., "code already exists" when it exists in another tenant)
  • Security audit failures

When Disabling is Justified (Rare)

Only disable multi-tenancy with explicit justification comment:

// ✅ JUSTIFIED: License plate numbers must be globally unique across all tenants
// to ensure physical warehouse operations don't conflict between tenants sharing facilities.
// Approved by: [Name] on [Date]
using (_dataFilter.Disable<IMultiTenant>())
{
    var existsGlobally = await _licensePlateRepository.AnyAsync(
        lp => lp.LicensePlateNumber == input.LicensePlateNumber && !lp.ShippedOut);
}

Multi-Tenancy Security Checklist

  • No _dataFilter.Disable<IMultiTenant>() without documented justification
  • Cross-tenant uniqueness checks are truly required (not accidental)
  • Error messages don't reveal other tenants' data
  • Audit logging captures cross-tenant operations
  • Unit tests verify tenant isolation

Common Vulnerability Patterns

Missing Authorization

// VULNERABLE
public async Task<PatientDto> GetAsync(Guid id)
{
    return await _repository.GetAsync(id);
}

// SECURE
[Authorize({ProjectName}Permissions.Patients.Default)]
public async Task<PatientDto> GetAsync(Guid id)
{
    return await _repository.GetAsync(id);
}

Information Disclosure in Errors

// VULNERABLE
catch (Exception ex)
{
    return BadRequest(ex.ToString()); // Exposes stack trace
}

// SECURE
catch (Exception ex)
{
    _logger.LogError(ex, "Error processing request");
    throw new UserFriendlyException("An error occurred");
}

Insecure Direct Object Reference

// VULNERABLE: Any user can access any patient
[HttpGet("{id}")]
public async Task<PatientDto> Get(Guid id)
{
    return await _service.GetAsync(id);
}

// SECURE: Verify user can access this patient
[HttpGet("{id}")]
public async Task<PatientDto> Get(Guid id)
{
    var patient = await _service.GetAsync(id);
    if (!await CanAccessPatient(patient))
        throw new UnauthorizedAccessException();
    return patient;
}

References

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

04

需要参考平台分布和安装热度时

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

补充不同宿主或平台的使用分布数据

能力 5

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

Cursor

30.09%
按下载量换算36

Claude Code

20.09%
按下载量换算24

github-copilot

18.17%
按下载量换算22

mcpjam

10.79%
按下载量换算13

crush

7.45%
按下载量换算9

cline

2.99%
按下载量换算4

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

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。

来源信息

继续浏览同类 Skills