The Agent Skills Directory

[SAFE]: The skill provides legitimate instructions for configuring CodeQL analysis. It follows official documentation and security best practices for GitHub Actions.
[EXTERNAL_DOWNLOADS]: References to external downloads are limited to official GitHub release pages (github.com/github/codeql-action/releases) for the CodeQL bundle. These are well-known, trusted sources.
[CREDENTIALS_UNSAFE]: The instructions correctly use environment variables (GITHUB_TOKEN) and GitHub secrets for authentication, avoiding hardcoded credentials.
[COMMAND_EXECUTION]: Command examples provided (such as codeql database create and tar xf) are standard administrative and operational tasks required for tool installation and execution.
[PROMPT_INJECTION]: There are no detected prompt injection patterns. Terms like 'CRITICAL' and 'IMPORTANT' are used appropriately within the context of security alert severity levels and configuration steps.

codeql