Security audit
Security checks for vulnerabilities and agentic risk
The skill is a coherent local Git report generator, but crafted report parameters could be inserted into Bash commands and run unintended shell code.
Review before installing. This skill does not appear to exfiltrate data or persist itself, but it will inspect local repository metadata and work-in-progress state. Avoid using untrusted or oddly formatted values for --author or day counts until the command templates add strict validation or structured argument passing, and be cautious in repositories with sensitive commit messages, branch names, stash entries, or contributor emails.
SKILL.md:57Potential Shell Command Injection Through Unvalidated Report Parameters
SKILL.md:43Unnecessary Collection of Contributor Email Addresses
Automatically choosing the output language based on commit message language changes behavior using repository content rather than explicit user preference. While not directly enabling code execution or privilege escalation, it can surprise users, leak linguistic signals from commit history into the response, and reduce predictability for a tool that processes potentially sensitive local development data.
The natural-language triggers are broad enough that ordinary requests like '帮我写今天的站会' or '生成一下本周的周报' could invoke the skill without an explicit command. Because this skill reads local git history, branch state, stash data, and uncommitted diffs, unintended activation can expose repository-derived content or summarize sensitive work context when the user did not clearly intend to grant that access.
The skill description says it uses local git commands but does not clearly warn that it will inspect commit history, branch state, stash entries, author identity, and uncommitted changes. Because the workflow includes git status, git diff, git diff --cached, and contributor enumeration, a user may unknowingly disclose sensitive local development details to the agent output.
The auto-trigger conditions are broad enough to match common requests like writing a daily update or weekly report, which can cause the skill to activate without the user realizing it will inspect local git history, branches, stashes, and uncommitted changes. In this skill’s context, unintended activation exposes sensitive repository metadata and work-in-progress details, so the issue is more dangerous than a generic false trigger.
The title and description indicate the skill's example output is presented in Chinese only, implying a fixed output language. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale restriction is explicitly justified.
The instruction '语言跟随项目' requires output to follow the language detected in commit messages, with no user opt-in or override. This imposes a locale/language choice derived from repository content instead of allowing the user to choose their preferred output language.
This markdown example presents headings and explanatory text in both Chinese and English, which can implicitly impose a language/locale format on users without opt-in. Under the language/locale policy check, this is a natural-language policy concern because no rationale or user choice is provided for the mixed-language output style.
No suspicious patterns detected.