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
prompt-engineering
Security checks for vulnerabilities and agentic risk
Overview
This is a Chinese prompt-engineering guide that changes response style and reasoning workflows but does not install code, persist data, or hide unrelated behavior.
Install this if you want a Chinese prompt-engineering and reasoning-mode guide. Be aware it may make agents use structured reasoning templates and, for ReAct-style tasks, choose tools such as search, file reading, exec, or browser when the task calls for outside information.
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
- 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
Findings (9)
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
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
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
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 86% confidence
- Finding
The manifest description is written as a Chinese-only skill description and the document consistently instructs behavior in Chinese, with no indication that users may choose their preferred language. This creates a natural-language locale policy concern because the skill appears to force a specific language without opt-in or justification.
- Content
Vague Triggers
Medium
- Category
- Not specified by scanner
- Confidence
- 90% confidence
- Finding
Several triggers are broad, generic, or overlap with common prompting terminology, which can cause unintended invocation of this skill outside the author's intended context. In an agent ecosystem, accidental activation may route tasks into a prompt-engineering workflow that changes outputs, reasoning style, or tool selection unexpectedly, creating reliability and policy-enforcement risks.
- Content
Vague Triggers
Medium
- Category
- Not specified by scanner
- Confidence
- 90% confidence
- Finding
The automatic mode-selection priority is underspecified relative to the later guidance that the default should be direct answering unless necessary, so agents may activate prompting modes inconsistently. In practice, this ambiguity can lead to unintended reasoning-mode invocation, extra tool calls, and unpredictable behavior across deployments or prompts.
- Content
Vague Triggers
Medium
- Category
- Not specified by scanner
- Confidence
- 94% confidence
- Finding
The explicit trigger phrases are broad, common-language terms like “搜索”, “计划”, and “一步一步”, which can cause the skill to activate reasoning modes in situations the user did not intend. In an agent setting, this increases the risk of unnecessary tool use, unexpected disclosure of reasoning structure, or costly multi-step behavior being triggered by ordinary phrasing.
- Content
Natural-Language Policy Violations
Low
- Category
- Not specified by scanner
- Confidence
- 88% confidence
- Finding
The guide forces a single language/locale in its natural-language instructions and examples, but does not state that Chinese is optional, user-selected, or required for a region-specific purpose. This may violate language or locale policy where users should be given a choice unless the restriction is documented and justified.
- Content
Static analysis
No suspicious patterns detected.