Back to skill

Security audit

企业-出海产能与机构外扩监测

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed public-research workflow for analyzing companies' overseas expansion and does not show hidden execution, persistence, credential use, or destructive behavior.

Before installing, users should expect Chinese-language outputs and should confirm the company, region, and time range when a query is ambiguous. Treat the generated strategic conclusions as research analysis that depends on public source quality, and verify important business decisions against the cited primary sources.

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 (2)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill instructions and expected operation are entirely in Chinese, with no user language negotiation or stated locale restriction. This can cause misinterpretation of user intent, incorrect outputs, or silent exclusion of non-Chinese users, especially when the skill is invoked by broad triggers and relies on precise multilingual source validation.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The trigger phrases are very broad and can match ordinary requests about a company's strategy, overseas activity, or globalization plans without clear scope limits. This creates unintended activation risk, which may route users into a highly specialized workflow that performs expansive web research and inference on sensitive corporate topics, increasing the chance of irrelevant collection, overreach, or misleading analysis.

Content

No source excerpt is available for this finding.