The Agent Skills Directory

[SAFE]: The skill is strictly instructional and does not contain code, shell commands, or network operations. No security threats across the 10 monitored categories were identified.
[PROMPT_INJECTION]: Instructions are focused on performance metrics and optimization strategies without any attempts to bypass safety protocols or manipulate the underlying agent behavior.
[DATA_EXFILTRATION]: There is no evidence of sensitive data access or external transmission mechanisms. All optimization actions described are local to the agent's prompt and context management.

accelerate