The header says no keys are required, but the README explicitly includes `gh auth login` and `gh auth status`, indicating practical use of the GitHub CLI typically relies on GitHub account authentication or a token. There is no evidence that the skill itself harvests credentials or sends them to a third party; the main concern is misuse of locally configured `gh` credentials if they have broad scope.

The skill is centered on using `gh` for issues, PRs, CI, releases, and `gh api` queries, which means real use will send requests and possibly user-provided text to GitHub's official services. No other remote endpoints are declared, and there is no sign of data being exfiltrated to unknown hosts.

The system flags this as prompt-only; the material simply explains how to use existing `gh` commands and does not include executable scripts, installers, or autonomous execution logic. While it contains command-line examples, the skill itself does not add a code execution surface.

There is no indication that the skill itself can directly read or write arbitrary local files, system directories, or private datasets; the `--body-file /tmp/...` examples are merely CLI usage examples and do not grant file access by themselves. Effective data access depends mainly on the repository permissions already granted to the GitHub CLI.

The source is GitHub, and the system marks it as open-source with very high community adoption (about 377k stars), both strong risk-reducing signals. Although the license is unspecified and maintenance status is unknown, the available evidence still suggests good auditability and community trust, with no red flags such as closed-source exfiltration or suspicious prompt injection.