The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill uses perl -c and perl -wc for syntax validation in both the instructions and the provided validation script. In Perl, the compile-check flag (-c) executes code contained within BEGIN, UNITCHECK, CHECK, and INIT blocks. If an untrusted or malicious script is processed, these blocks will be executed on the host system, leading to arbitrary code execution during a process intended to be a static check.
[COMMAND_EXECUTION]: The skill utilizes the Bash tool to run several command-line utilities, including perl, grep, podchecker, and perlcritic. These tools operate directly on user-provided file paths and content.
[INDIRECT_PROMPT_INJECTION]: The skill creates a surface for indirect injection by ingesting external Perl scripts and processing them with tools that can execute parts of the input.
Ingestion points: The skill reads Perl scripts provided by the user or identified via Glob and Read operations.
Boundary markers: There are no explicit delimiters or warnings to the agent regarding the handling of embedded instructions or the risks of the Perl compilation phase.
Capability inventory: The agent has access to the Bash tool to execute compilers and static analysis tools on the ingested content.
Sanitization: The skill does not perform any pre-processing or sanitization of the script content to remove or disable executable blocks before invoking the Perl compiler.

perl-validate