The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read and parse several project files, such as composer.json, composer.lock, and rector.php. These files represent untrusted external inputs that could potentially contain malicious instructions designed to subvert the agent's behavior.
Ingestion points: composer.json, composer.lock, vendor/composer/installed.json, rector.php, phpunit.xml, phpstan.neon, bootstrap/app.php, and app/Http/Kernel.php.
Boundary markers: The skill does not provide specific delimiters or instructions to ignore potential commands embedded within these project files.
Capability inventory: The skill facilitates the execution of various shell commands including composer, vendor/bin/rector, php artisan, and vendor/bin/phpstan based on the data gathered from the project files.
Sanitization: No explicit sanitization or validation of the content of these files is mandated before the agent processes them or uses the derived information to execute commands.
[COMMAND_EXECUTION]: The skill guides the agent to execute multiple CLI tools (composer, rector, php artisan, pint, phpstan). While these are standard development tools for the PHP/Laravel ecosystem, they are executed in the local environment and their specific actions are determined by the contents of the project's configuration files.

laravel-update-with-rector