The Agent Skills Directory

[COMMAND_EXECUTION]: The skill exposes the execute_terminal_command and execute_run_configuration tools, which allow the agent to run arbitrary shell commands and defined IDE tasks. These capabilities are intended for development workflows such as running tests, building projects, and executing Artisan or Symfony console commands within the local development environment.
[INDIRECT_PROMPT_INJECTION]: The skill operates by having the agent read and analyze project-specific files (e.g., composer.json, composer.lock, and PHP source files), which presents an attack surface for indirect prompt injection. A malicious file within a project could theoretically attempt to influence the agent's logic through embedded instructions.
Ingestion points: The agent utilizes read_file, get_file_text_by_path, and get_composer_dependencies to ingest content from the project directory.
Boundary markers: No specific boundary markers or instruction-ignoring delimiters are defined in the provided documentation for handling the content of these files.
Capability inventory: The skill possesses the ability to execute terminal commands, create or modify files, and perform refactoring operations.
Sanitization: The instructions do not prescribe specific sanitization or validation of the ingested content before it is processed by the agent.

php-project-guide