The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): The skill includes a Python script scripts/pattern_scanner.py for text analysis. Analysis of the code shows it uses standard regular expressions and does not utilize unsafe functions like eval(), exec(), or subprocess to run external commands.
[DATA_EXFILTRATION] (SAFE): The script does not contain any network-related code (no requests, urllib, or socket operations). It processes text locally and outputs results to the console.
[REMOTE_CODE_EXECUTION] (SAFE): No external dependencies are required or downloaded. The script relies entirely on the Python standard library.
[PROMPT_INJECTION] (SAFE): The instructions in SKILL.md are focused on linguistic analysis and do not attempt to bypass agent safety filters or extract system prompts.
[INDIRECT_PROMPT_INJECTION] (LOW): While the skill ingests untrusted user text for analysis (Ingestion point: pattern_scanner.py), the capability inventory is restricted to regex-based counting and JSON reporting. There is no execution path for the analyzed content to influence the agent's core logic or access protected resources.

ai-pattern-detection