Security audit
Security checks for vulnerabilities and agentic risk
This Git helper is not clearly malicious, but it asks for command and file-changing authority while giving inconsistent and overly broad instructions.
Review this skill before installing. It appears to be a low-quality Git automation prompt rather than malware, but it can guide an agent to run commands and modify repositories, and its API/credential language is vague. Use it only in repositories you are prepared to change, avoid giving it unrelated filesystem or credential access, and require confirmation before rebase, history recovery, push, or any bulk file changes.
The description includes broad language about improving efficiency, automation, batch processing, and workflow optimization beyond narrowly scoped Git tasks. In an agent ecosystem, this can cause the skill to be invoked for overly general requests and then use read/write/exec capabilities in contexts the user did not clearly intend.
The document claims that only whitelisted commands are executed, but it does not define any whitelist, enforcement mechanism, or parameter constraints while exposing exec-based operation. This creates a false sense of safety and leaves room for arbitrary command execution behavior if an agent follows the broad instructions literally.
The Markdown describes command execution, file handling, and external API use, but does not present a clear, up-front warning that these actions may modify repositories, overwrite files, expose local data, or trigger network transmission. For a skill with exec/read/write capabilities, insufficient user warning increases the chance of unsafe or unintended side effects.
The skill documentation claims the skill is pure Markdown/natural-language driven, but the manifest declares exec, write, and read tools and other sections describe command execution and API calls. This mismatch can mislead users or downstream policy systems about the actual capability and risk surface, causing the skill to be granted broader trust than warranted.
The display name, summaries, and description are presented in a mixed Chinese-English form, but the document does not state language options or let the user choose a preferred language. This can conflict with language/locale policy expectations when a skill effectively defaults to a particular bilingual presentation without explicit user opt-in.
No suspicious patterns detected.
