Security audit
Security checks for vulnerabilities and agentic risk
This is a generic coding helper with disclosed local read and command execution access; it has documentation gaps, but no hidden payload, persistence, or exfiltration behavior was found.
Install only if you want a coding assistant that may read project files and run local commands. Review commands before execution, especially install, deployment, deletion, or network commands, and do not export an API_KEY unless the agent platform or service you are actually using requires it.
The invocation criteria are very broad, covering general coding, debugging, testing, and deployment scenarios without narrow triggers or exclusions. This makes the skill likely to activate for many common developer requests, potentially exposing users to exec-capable behavior in situations where they may not expect environmental side effects.
The skill enables exec while describing a broad coding workflow and does not define clear command boundaries, approval requirements, or sandbox expectations. In an agent context, this can lead to unintended execution of project-supplied or user-influenced shell commands, increasing the risk of local file modification, data exposure, or destructive actions.
The skill advertises exec capability but does not clearly warn users that it may run commands in their local environment. Without an explicit warning, users may provide prompts or repositories assuming analysis-only behavior, while the skill can perform side-effecting actions such as installing packages, modifying files, or invoking tools.
The natural-language description is presented in a mixed Chinese/English form and does not state that outputs will follow the user's preferred language or locale. This can create a language-policy issue if the skill defaults to a specific language presentation without explicit user opt-in or choice.
Lines L135-L138 state that the required LLM API is provided by the agent and leave API key configuration blank, but L144-L148 then instruct users to export an API_KEY. This creates a documentation-level contradiction about whether external credentials are needed for the skill.
No suspicious patterns detected.
