The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes external tools and commands such as php artisan for Laravel validation and the fuse-ai-pilot:sniper utility for post-modification consistency checks. These operations are performed within the local development context as part of the primary refactoring workflow.
[PROMPT_INJECTION]: The skill analyzes existing codebase files to map dependencies and plan migrations, creating a surface for indirect prompt injection from potentially malicious source code.
Ingestion points: Reads and parses files from the project directory during the exploration phase.
Boundary markers: The process includes a mandatory human-in-the-loop checkpoint where the agent must present a migration map and wait for explicit user approval before performing any modifications.
Capability inventory: The skill can move files, create new directories, modify source code namespaces, and execute local validation commands.
Sanitization: There is no explicit sanitization of codebase content mentioned, though the workflow includes validation steps to detect broken classes or imports.

modularize