Security audit
Security checks for vulnerabilities and agentic risk
This skill is a narrow GitHub repository lookup helper, with some overbroad trigger wording and public-contact-data exposure to be aware of.
Install this only if you want an agent to use the UAPI GitHub repository lookup endpoint. Because the trigger list includes broad terms such as 'repo', review invocations where the user did not clearly ask for a GitHub repository lookup. Treat returned collaborator, maintainer, and email fields as public personal data and avoid bulk harvesting or unsolicited outreach.
The skill description includes trigger terms like "repo" and "github repo" that are overly broad and likely to match many unrelated user requests. This can cause unintended invocation of the skill, leading the agent to route queries to the GitHub repository endpoint when the user did not explicitly ask for that API action.
The common keywords section explicitly lists ambiguous phrases including "github repo" and especially "repo", which are common everyday terms with many meanings. In an agent-routing context, this broad matching surface increases the chance of false activation and unintended external API usage, which can confuse users and mis-handle tasks.
The skill explicitly advertises returning collaborators and inferred maintainers, including their public email addresses if available, without any privacy warning, minimization guidance, or purpose limitation. Even if the emails are publicly accessible upstream, aggregating and surfacing them in a convenience API increases the likelihood of misuse such as scraping, unsolicited contact, or targeted phishing.
The trigger keyword list includes very broad terms such as "repo" and "github repo," which can match ordinary user speech far beyond the narrowly intended action of querying a repository via a single endpoint. In an agent routing context, this can cause incorrect skill selection, unnecessary external API calls, and unintended disclosure of user prompts or parameters to a third-party service.
The file presents both Chinese and English phrasing, but the example guidance includes English trigger language as a prescribed invocation style without stating that language choice is optional. This can be read as preferring or imposing a specific language mode rather than explicitly allowing the user's preferred language.
No suspicious patterns detected.
