The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes use of standard Laravel CLI utilities, including php artisan for managing migrations, routes, and queues, as well as pint for code styling and pest for test execution. These are expected tools within the PHP/Laravel development ecosystem.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to take user-supplied application requirements and database schemas to generate code. While this constitutes an attack surface for indirect prompt injection, the skill specifically mandates security constraints to mitigate common vulnerabilities like SQL injection and unvalidated input.
Ingestion points: The 'Analyse requirements' step in the core workflow processes external developer input.
Boundary markers: No explicit delimiters for external input are defined in the instructions.
Capability inventory: The agent can write PHP source files and execute CLI tools like php artisan.
Sanitization: The instructions include explicit 'MUST NOT DO' rules against raw queries and skipping validation, which helps ensure the generated code remains secure.

laravel-specialist