Tp2
High
- Category
- MCP Tool Poisoning
- Confidence
- 85% confidence
- Finding
Mixing characters from multiple Unicode scripts in a single identifier is a common technique to create visually ambiguous tool names.
- Content
Security audit
Qzone
Security checks for vulnerabilities and agentic risk
Overview
This is a Markdown-only QQ Zone guidance skill; it has some scope wording drift but no hidden execution, data access, persistence, or credential handling.
Before installing, understand that this is broader than a backup/privacy helper: it also gives QQ Zone content and engagement advice. It does not appear to run code or access accounts, but users should still handle deletion and third-party export tools carefully as the skill itself warns.
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
- MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
- 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)
Description-Behavior Mismatch
Medium
- Category
- Not specified by scanner
- Confidence
- 95% confidence
- Finding
The manifest describes a narrow helper for legacy-data export/backup and privacy configuration, but the body expands into broad content-operations, engagement, and growth advice. This scope drift can cause agents or users to invoke the skill for purposes not disclosed in metadata, weakening trust boundaries and enabling unexpected behavior selection in larger agent systems.
- Content
Intent-Code Divergence
Medium
- Category
- Not specified by scanner
- Confidence
- 89% confidence
- Finding
The role section claims expertise in content operation, social interaction strategy, and fan growth, which materially exceeds the manifest's stated backup/privacy/legacy-support purpose. In agentic environments, this mismatch may lead to overbroad delegation, user surprise, or routing of social-engineering-adjacent tasks to a skill that was presented as a maintenance/privacy assistant.
- Content