The Agent Skills Directory

[SAFE]: The skill is a comprehensive educational resource for secure Perl programming, covering taint mode, input validation, and secure system calls.
[COMMAND_EXECUTION]: Potential destructive commands (e.g., rm -rf /) were identified in comments and anti-pattern sections. These are used strictly as illustrative examples of command injection vulnerabilities that developers must avoid, rather than instructions for the agent to execute.
[REMOTE_CODE_EXECUTION]: The skill accurately identifies eval with user input as a critical risk and explicitly categorizes it as an anti-pattern to be avoided.
[PROMPT_INJECTION]: The skill contains no attempts to override agent behavior or bypass safety guidelines. Its instructions are focused on providing security expertise to the user.
[SQL_INJECTION]: The skill provides correct and detailed guidance on using DBI placeholders and parameterized queries to prevent SQL injection attacks.

perl-security