Security audit
Security checks for vulnerabilities and agentic risk
This skill creates and maintains project guide documentation, with file writes and project-doc reads that are disclosed and aligned with that purpose.
Install this if you want a Chinese-language guide documentation workflow. Review the target --root before running scripts, avoid --force unless overwriting docs is intended, and remember that future agents may read guide notes and pitfalls as project context.
该脚本的实际主功能是初始化项目文档骨架，这与声明中的“初始化模式”部分一致；但声明将技能描述为一个支持初始化、更新、上下文恢复三种模式的完整文档体系维护工具，而代码中只有 init_guide.py，且只接受项目名、类型、根目录和 force 覆盖参数，执行创建/覆盖 Markdown 模板文件。代码没有读取已有文档并追加更新日志、踩坑记录或同步版本号的逻辑，也没有分析项目内容、恢复上下文或为新对话生成项目摘要的能力。因此描述明显高于实际实现，属于功能性不匹配。
The declared description presents a broader skill for creating and maintaining a full standardized project guide system across its lifecycle, including one-click initialization and AI context recovery in new conversations. The supplied code only implements update-oriented maintenance actions for an already existing guide structure. It explicitly exits if guide/ does not exist and tells the user to run a separate init_guide.py, showing initialization is not included here. It also contains no functionality related to restoring conversational/project context for AI use. While parts of the declared update mode are represented (changelog, pitfalls, version update, file checklist comparison), the implemented behavior is materially narrower than the declared purpose, so this is a description-behavior mismatch.
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
## 1. 设计理念
### 1.1 为什么做这个项目？
<!-- 项目的起源、灵感、要解决的问题 -->
### 1.2 核心设计原则
<!-- 项目遵循的设计原则（3-5条） -->
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
## 1. 设计理念
### 1.1 为什么做这个项目?
<!-- 项目的起源、灵感、要解决的问题 -->
### 1.2 核心设计原则
<!-- 项目遵循的设计原则(3-5条) -->
This template explicitly includes sections for 'system prompt' content inside an AI system design document. In an agent-skill ecosystem, embedding prompt templates can become dangerous if downstream tooling later ingests generated documentation as live instructions, creating a prompt-injection or instruction-confusion path between documentation and execution contexts.
<!-- system prompt 模板 -->
### 行动 Prompt（action）
<!-- system prompt 模板 -->
## LLM 集成
The skill instructs reading and writing project files but does not declare any explicit tool scope or permission boundaries. That makes file access implicit and harder to audit, increasing the risk of unintended modification of arbitrary project paths when the skill is triggered.
The context-recovery condition is ambiguous because it suggests activation at the start of a new conversation without a clear user request. That can prompt automatic reading of project documentation and unintended access to potentially sensitive internal files, expanding exposure beyond the user's immediate intent.
The context-recovery condition is ambiguous because it suggests activation at the start of a new conversation without a clear user request. That can prompt automatic reading of project documentation and unintended access to potentially sensitive internal files, expanding exposure beyond the user's immediate intent.
The skill directs file creation and document updates but does not prominently warn the user that it will modify project files. Without explicit notice and confirmation, users may trigger persistent changes unintentionally, especially given the broad activation phrases and maintenance-oriented workflow.
The file title and core instruction text are entirely in Chinese and position the document as the authoritative reference for AI behavior during initialization/update modes. This imposes a specific language/locale expectation on the AI without offering user choice or documenting a justified region-specific constraint.
This markdown template appears to require Chinese for headings, instructions, and examples, but it does not provide any user opt-in or alternative language option. Under the policy, forcing a specific language without user choice is a natural-language policy violation unless the locale restriction is explicitly justified.
This markdown template is entirely written in Chinese, including all headings and usage instructions, which imposes a specific language on downstream skill documentation. The policy explicitly disallows forcing a language or locale unless the skill offers user opt-in or clearly documents a justified regional constraint, neither of which is present here.
Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
cd /path/to/project
nohup node server.js > log/server.log 2>&1 &
echo "服务已启动，PID: $!"
The template’s headings and instructions are entirely in Chinese, which can impose a specific language/locale on downstream skill outputs. The policy allows locale constraints only when they are optional, user-selected, or clearly justified, which is not indicated here.
The template title and surrounding section labels are written in Chinese, which implicitly prescribes a specific language for generated project documentation. Because the file does not offer a language choice or explain that the template is intended only for a Chinese-speaking context, this can violate a language/locale policy requiring user opt-in.
The entire skill template is written as a Chinese-only document, including headings, field labels, and examples, with no indication that users may choose another language or that the template is intended only for a Chinese-language context. Under the policy for natural-language violations, forcing a specific language without opt-in is reportable.
No suspicious patterns detected.