The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains structured instructions to guide the model's reasoning process and output format. No attempts to bypass safety guidelines, extract system prompts, or override core instructions were identified.
[COMMAND_EXECUTION]: While the skill metadata lists 'Bash' as an allowed tool, the instructions themselves do not contain any shell commands or logic that would execute arbitrary scripts or modify the host system.
[DATA_EXFILTRATION]: The skill does not include any network operations or patterns associated with reading sensitive local files and transmitting them to external domains.
[EXTERNAL_DOWNLOADS]: No remote URLs or external package dependencies are referenced within the skill code.

reasoning-tracer