Back to skill

Security audit

邮储银行易企收

Security checks for vulnerabilities and agentic risk

Overview

This skill only gives instructions for opening disclosed 易企收 demo or login pages in a visible browser, with no code, persistence, credential collection, or hidden behavior.

Install only if you want an agent to open 易企收 demo or login pages for you. Because the triggers are broad and the site is payment-related, review the destination before entering credentials or business data, and ask the agent for confirmation when your request is informational only.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill registers on broad natural-language phrases such as '打开易企收', '管理缴费信息', and '了解缴费功能', which can overlap with ordinary user discussion and cause unintended invocation. Because the skill can automatically open external banking/payment-related pages, accidental triggering may launch browser actions without sufficiently specific user intent.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The keyword-based intent rule treats vague words like '了解', '查看', or '演示' as sufficient to decide navigation behavior, but it lacks boundaries, exclusions, or confidence thresholds. This can misclassify conversational requests and automatically send users to demo or production-related pages, increasing the chance of unwanted external actions in a sensitive payment context.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The workflow directs the agent to automatically open official or demo URLs based on inferred intent, but it does not require a clear notice or user confirmation before launching the browser. While the destinations appear legitimate, silent navigation can still surprise users, create privacy concerns, and make accidental interaction with production login pages more likely.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.