The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides various standard development commands in SKILL.md and composer-scripts.json. These involve standard tools like bun, composer, php artisan, and pint. All commands are within the expected scope of setting up and maintaining a Laravel project.
[EXTERNAL_DOWNLOADS]: The skill references standard, well-known GitHub Actions (actions/checkout, shivammathur/setup-php, oven-sh/setup-bun). It also instructs the user to install laravel-wayfinder from a GitHub repository belonging to earendil-works. These are documented as part of the intended setup process and do not include any non-standard or suspicious execution patterns.
[PROMPT_INJECTION]: No prompt injection patterns, bypass markers, or attempts to override agent behavior were detected in the instructions or metadata.
[DATA_EXFILTRATION]: No patterns associated with credential harvesting or sensitive data exfiltration were found. The workflows use persist-credentials: false as a security best practice.
[COMMAND_EXECUTION]: The pre-commit hook in references/pre-commit executes lint-staged. This is a common and expected practice for enforcing code style and linting rules before a commit is finalized.

project-bootstrap