Improper Access Control
The product does not restrict or incorrectly restricts access to a resource from an unauthorized actor.
Access control involves determining which subjects can access which objects. When access control is implemented incorrectly, it can lead to unauthorized access to sensitive data or functionality.
如何修复此漏洞
基于 4 条 Shoulder 检测规则的 Improper Access Control 预防策略。
Validate tool inputs against strict schemas and use an allowlist for permitted tools
- func handleToolCall(toolCall ToolCall) (interface{}, error) { - return tools[toolCall.Name](toolCall.Arguments) + var toolRegistry = map[string]ToolConfig{ + "search": {Handler: searchHandler, Validator: validateSearch, Permission: "read"}, + "weather": {Handler: weatherHandler, Validator: validateWeather, Permission: "read"}, + } + + func handleToolCall(userPerms map[string]bool, toolCall ToolCall) (interface{}, error) { + config, ok := toolRegistry[toolCall.Name] + if !ok { + return nil, fmt.Errorf("unknown tool: %s", toolCall.Name) + } + if !userPerms[config.Permission] { + return nil, fmt.Errorf("permission denied for tool: %s", toolCall.Name) + } + args, err := config.Validator(toolCall.Arguments) + if err != nil { + return nil, fmt.Errorf("invalid arguments: %w", err) + } + return config.Handler(args) }
Validate tool inputs against schemas and use allowlists for permitted tools
- for (const toolCall of response.tool_calls) { - const fn = tools[toolCall.function.name]; - const result = await fn(JSON.parse(toolCall.function.arguments)); + const allowedTools = new Set(['search', 'calculate', 'getWeather']); + + for (const toolCall of response.tool_calls) { + if (!allowedTools.has(toolCall.function.name)) { + throw new Error('Unknown tool'); + } + const validate = ajv.compile(toolSchemas[toolCall.function.name]); + const args = JSON.parse(toolCall.function.arguments); + if (!validate(args)) throw new Error('Invalid arguments'); + await tools[toolCall.function.name](args); }
Define NetworkPolicy resources to restrict pod-to-pod traffic and enforce network segmentation
- apiVersion: apps/v1 - kind: Deployment - metadata: - name: web - spec: - replicas: 3 - template: - spec: - containers: - - name: web - image: nginx:1.25 + apiVersion: networking.k8s.io/v1 + kind: NetworkPolicy + metadata: + name: web-policy + spec: + podSelector: + matchLabels: + app: web + policyTypes: + - Ingress + ingress: + - from: + - podSelector: + matchLabels: + role: frontend + ports: + - port: 80
Use Pydantic for tool input validation and maintain a strict allowlist for permitted tools
- def handle_tool_call(tool_call): - name = tool_call.function.name - args = json.loads(tool_call.function.arguments) - return tools[name](args) + from pydantic import BaseModel, Field + + class SearchArgs(BaseModel): + query: str = Field(max_length=100, pattern=r'^[a-zA-Z0-9\s]+$') + + ALLOWED_TOOLS = {'search_products': SearchArgs, 'get_weather': WeatherArgs} + + def handle_tool_call(tool_call): + name = tool_call.function.name + if name not in ALLOWED_TOOLS: + raise ValueError(f'Unknown tool: {name}') + schema = ALLOWED_TOOLS[name] + args = schema.parse_raw(tool_call.function.arguments) + return handlers[name](args)
查找代码中的漏洞
使用Shoulder扫描代码中的Improper Access Control模式。 4 规则.
# Scan with Shoulder CLI npx @shoulderdev/cli trust --cwe=284 # Or scan entire project npx @shoulderdev/cli trust .
检测规则 (4)
代码审查中需要关注的内容
这些模式表明潜在的Improper Access Control漏洞。在代码审查和安全审计中注意查找。