The Agent Skills Directory

[SAFE]: The skill is an instructional resource for secure Perl development. It does not contain any functional malicious code or requests for escalated privileges.
[COMMAND_EXECUTION]: The guide demonstrates the risks of shell injection through instructional examples, such as using rm -rf / as a payload in a vulnerable string-form system call or 2-argument open call. It provides safe alternatives using list-form syntax to prevent such attacks.
[DATA_EXFILTRATION]: The skill identifies common ingestion points for untrusted data including environment variables, command-line arguments, and STDIN. It provides robust patterns for sanitization using Perl's taint mode (-T) and regular expression-based allowlisting to ensure data is safe before use.
[REMOTE_CODE_EXECUTION]: The guide warns against the use of eval with user-provided strings, categorizing it as a critical vulnerability (anti-pattern) and offering secure design patterns instead.
[SAFE]: Static analysis triggers on destructive commands like rm -rf / were evaluated and found to be false positives as they are contained within comments and code blocks demonstrating vulnerable patterns to avoid.

perl-security