The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill presents a potential attack surface by ingesting and processing untrusted data from the local PHP project environment (source files and configuration) to perform manual and automated fixes.
Ingestion points: Files identified and read via find_files_by_glob and get_file_text_by_path during the Discovery and Fix phases.
Boundary markers: Absent; the instructions do not explicitly warn the agent to treat ingested source code strictly as data.
Capability inventory: The skill possesses file-writing capabilities via replace_text_in_file and apply_quick_fix across the project directory.
Sanitization: No input sanitization or validation of the content read from files is performed before it is presented to the agent for decision-making.

upgrade-php