Back to skill

Security audit

命题工坊

Security checks for vulnerabilities and agentic risk

Overview

This is a Chinese-language contest-problem-writing skill with no code execution, data access, persistence, or hidden behavior.

Install this if you want Chinese NOI/CSP/OI problem-generation guidance. Users working in other languages or seeking general coding help should be aware it may steer responses toward Chinese OI-style problem authoring.

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
90% confidence
Finding

The skill metadata and instructions are fully Chinese-centric and do not offer a user-language negotiation path, which can force output language/context regardless of the user's preferences. This is not a code-execution risk, but it can degrade usability, cause incorrect routing or non-compliant responses, and make the skill behave unexpectedly in multilingual environments.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger scope is overly broad because it activates for essentially any algorithm-competition-related request, which can cause the skill to hijack user intent and override more appropriate general-purpose or safety-aware behaviors. In an agentic system, broad routing increases the chance of misclassification, unwanted prompt dominance, and accidental application outside the narrow intended NOI/CSP problem-setting context.

Content

No source excerpt is available for this finding.