Security audit
Security checks for vulnerabilities and agentic risk
The skill is a disclosed discussion-documentation workflow that writes Markdown records in the workspace and optionally performs user-requested Git versioning.
Install this only if you want technical discussions saved as files in your workspace. Review the generated documents before sharing the repository, and inspect Git status before using the optional version-recording flow because it may create commits and tags.
Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).
{
"id": "cleanup-notice",
"kind": "shell",
"script": "echo '' && echo '📁 tech-discussions/ 目录已保留在 workspace 中。' && echo ' 该目录可能包含你的讨论记录、决策文档等。' && echo ' 如需删除，请手动执行：rm -rf tech-discussions/' && echo ''",
"label": "卸载提示",
},
],
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
node_modules/
__pycache__/
*.pyc
.env
.venv/
venv/
.vscode/
The file consistently specifies names, triggers, workflow phrases, and user-facing instructions only in Chinese, which effectively forces a specific language for interaction. The policy allows locale constraints only when users are given a choice or when the restriction is explicitly justified as region-specific.
The skill directs the agent to automatically create and update files throughout the discussion lifecycle, but it does not require an explicit user-facing warning or confirmation before workspace writes occur. This can lead to unintended persistence of potentially sensitive discussion content, especially because writes are triggered by broad conversational cues like starting or ending a discussion.
The example shows the agent automatically creating and updating discussion files as soon as the workflow starts, but it does not disclose that filesystem state will be modified. In an agent environment, undocumented writes can surprise users, overwrite existing content, or normalize unsafe behavior where users do not realize persistent changes are happening.
These examples depict automatic generation of decision, architecture, and development-plan documents, plus creation of a code directory, without any consent or warning about repository changes. That can lead to unintended artifact sprawl, accidental overwrites, and unsafe expectations that the skill may freely modify project structure.
The Git example presents version recording and tag creation as a routine action but does not warn that repository state may be altered. Creating Git tags can affect release processes, CI/CD automation, and downstream consumers, so silently normalizing this behavior is risky in a repository-integrated skill.
Natural-language policy requires avoiding forced language or locale unless the skill offers a choice or clearly justifies the constraint. This file is entirely in Chinese and does not indicate that users can select another language or that the skill is intentionally limited to a Chinese-speaking context.
The document title is written entirely in Chinese, which signals that the skill template is intended to produce content in a fixed language. The file does not offer a language choice or explain why Chinese is required, which can violate language/locale policy constraints.
No suspicious patterns detected.