The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data in the form of PHP and JavaScript source code from WordPress themes and plugins. This creates an attack surface where malicious instructions hidden in the code being audited (e.g., in comments or large string literals) could attempt to influence the agent's behavior.
Ingestion points: The skill reads and searches through arbitrary files in the user's environment, such as functions.php, plugin.php, and other source files, often using shell tools like grep.
Boundary markers: There are no explicit instructions or delimiting markers to separate the reviewed code from the agent's internal instruction context, nor are there warnings to the agent to disregard instructions found within the data.
Capability inventory: The agent environment associated with this skill has terminal access (specifically for grep and file operations), file system read access, and potentially network access, which increases the potential impact of a successful injection.
Sanitization: The skill lacks mechanisms to sanitize or validate the content of the reviewed source code before it is processed by the LLM.

wp-performance-review