The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes standard shell commands for project reconnaissance and verification, including grep, cat, and vendor/bin/phpstan. These are local operations necessary for the skill's primary function of static analysis and error resolution.
[EXTERNAL_DOWNLOADS]: The skill recommends installing well-known community packages from official registries, such as larastan/larastan, phpstan/phpstan-symfony, and barryvdh/laravel-ide-helper. These are established tools within the PHP ecosystem used to improve analysis accuracy.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data including project source files and PHPStan error logs which could theoretically contain malicious instructions.
Ingestion points: Reads composer.json, phpstan.neon, AGENTS.md, and various PHP source files in the project.
Boundary markers: None explicitly specified for separating code content from instructions during ingestion.
Capability inventory: Has the ability to read/write files and execute shell commands (composer, php artisan, phpstan).
Sanitization: Relies on the user's manual review of proposed changes and uses a human-in-the-loop Question tool before performing significant actions like ignoring errors.

phpstan-fixer