Back to skill

Security audit

Prompt Optimizer by dxx

Security checks for vulnerabilities and agentic risk

Overview

This appears to be a disclosed Chinese-language prompt-rewriting utility with no hidden execution, credential access, network access, or persistence beyond installation.

Use this skill when you want Chinese-oriented prompt rewriting. Review the optimized prompt before using it, especially if you need a different language, output format, or programming language, because it may add defaults like Markdown, Python, and concise professional Chinese wording.

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 (8)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill markets itself as able to transform vague, generic user requests into optimized instructions, using examples like writing code, checking weather, and creating APIs. This creates an activation scope that overlaps heavily with ordinary conversation, increasing the chance the skill will trigger in contexts far beyond a narrowly defined purpose and unintentionally steer or rewrite user intent.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The supported task list spans many everyday intents such as chatting, summarizing, translation, weather lookup, coding, brainstorming, and math. Such broad coverage makes the skill effectively a general-purpose interceptor, which can cause over-application, prompt hijacking of unrelated requests, and unexpected modification of user inputs across routine interactions.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The enhancement string explicitly instructs the model to summarize '用简洁的中文概括', which imposes a specific language choice in natural-language behavior. The file does not indicate that this is optional, user-selected, or justified as a region-specific constraint.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The module docstring, CLI usage examples, and user-facing descriptions are written entirely in Chinese, and the optimization logic later injects Chinese instructions into prompts by default. There is no indication that users can choose another language or locale, which creates a language-policy issue for a general-purpose skill.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

When information is missing, the code appends fixed Chinese guidance such as '请用 Markdown 格式输出', '(请用 Python 实现)', and '请用简洁专业的风格'. This enforces a specific language/locale in generated prompt text regardless of the user's original language or preference.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

This Python test file contains natural-language expectations and console output entirely in Chinese, including prompts and task labels such as '翻译成英文' and expected categories like '写代码' and '通用任务'. That indicates the skill behavior is implicitly constrained to Chinese-language inputs/outputs, but there is no user opt-in or documented locale justification in the file.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

The natural-language instructions and usage guidance are presented entirely in Chinese, but the document does not state that the user can choose another language or that the locale restriction is intentional for a region-specific skill. This can be a language/locale policy issue because the skill appears to impose a fixed language by default without user opt-in.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

The module docstring and the implementation consistently describe functionality centered on prompt optimization, classification, and formatting. The CLI parser description broadens that intent to 'AI 任务处理器' (AI task processor), which suggests the skill performs AI task execution rather than just rewriting prompts, creating a documentation-to-code mismatch in the inline help text.

Content

No source excerpt is available for this finding.