The Agent Skills Directory

[SAFE]: The skill is composed solely of markdown files containing instructional templates and examples for logical code reasoning.
[PROMPT_INJECTION]: No instructions were found that attempt to bypass safety guardrails or override core agent behavior. The guidelines provided are limited to task-specific reasoning steps.
[DATA_EXFILTRATION]: There are no commands or instructions to access sensitive files or perform network requests.
[REMOTE_CODE_EXECUTION]: No external dependencies, scripts, or remote content are fetched or executed.
[OBFUSCATION]: The content is clear-text markdown without any hidden or encoded strings.
[INDIRECT_PROMPT_INJECTION]: Although the skill processes external code provided by users, the structured templates require the model to treat code as data rather than instructions, mitigating potential injection risks. The skill defines no tool-use capabilities.

reasoning-semiformally