The Agent Skills Directory

[SAFE] (SAFE): The file consists entirely of markdown documentation and templates for configuring AI agents. There are no active code components or scripts that could execute malicious commands.
[DATA_EXFILTRATION] (SAFE): No credentials, sensitive file paths, or network exfiltration patterns were detected. The templates use placeholders like 'agent-name-here'.
[PROMPT_INJECTION] (SAFE): The instructions provided are for legitimate role-play (e.g., 'You are a testing expert') and do not attempt to bypass safety filters or extract system prompts.
[REMOTE_CODE_EXECUTION] (SAFE): No remote downloads or dynamic execution patterns (eval/exec) are present in the documentation.

create-worker