Back to skill

Security audit

Prompt Master - 提示词工程精华版

Security checks for vulnerabilities and agentic risk

Overview

This is a prompt-writing reference skill with disclosed optional auto-loading and no executable, data-access, or destructive behavior.

Review the optional auto-load setting before enabling it, because it can make these prompt-writing rules influence sessions by default. The skill otherwise appears to be a low-risk Chinese prompt template reference; users who prefer another language or precise rule numbering may want revisions before relying on it.

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 (3)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger conditions are broad and include optional auto-loading at session start, which can cause the skill to inject prompt-shaping instructions into conversations without a clear user request. In a prompt-management skill, unclear invocation boundaries increase the chance of unintended behavioral override, instruction collision, and persistent interference with other security-relevant system or task prompts.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
87% confidence
Finding

The markdown presents all instructions, triggers, and examples exclusively in Chinese, which can effectively impose a language preference on users without any opt-in or justification. Under the language/locale policy, this is a natural-language policy concern unless the skill explicitly offers language choice or states a valid region-specific constraint.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

The document defines Rule 8 as a duplicate of '不知道就说不知道' at L114-L119, while later the rule-combination table maps 'AI回答太长' to rules 8 and 9 at L329 and the mnemonic maps rule 7 to '结论先行' and rules 8,9 to '简洁直接' at L342-L343. These internal references contradict the earlier rule definitions and could cause users to apply the wrong guidance.

Content

No source excerpt is available for this finding.