The Agent Skills Directory

[SAFE]: The skill is primarily instructional, providing patterns and logic for text parsing. It does not include executable scripts, shell commands, or network operations.\n- [DATA_EXFILTRATION]: The skill accesses local markdown files containing regulatory text within the internal/shared/regulations/ directory. This access is consistent with its purpose and does not target sensitive system or user configuration files. No network activity or data export patterns were found.\n- [PROMPT_INJECTION]: Although the skill processes external regulation files (indirect prompt injection surface), it lacks the capabilities (such as writing to files or executing commands) to be exploited. No direct prompt injection attempts or safety bypasses are present in the skill's instructions.\n- [DYNAMIC_EXECUTION]: The skill documentation includes pseudo-code for parsing logic. This code uses standard library functions and does not employ dangerous dynamic execution patterns such as eval() or exec() on untrusted data.

regulation-hierarchy-parser