The Agent Skills Directory

[SAFE]: The skill is a logic-only router designed to coordinate other skills (evp-onboarding, evp-craft, evp-brief). It uses standard file-check logic to determine the current state of a branding project.
[COMMAND_EXECUTION]: While the skill contains code-like blocks for state detection and logic tables, these are internal placeholders for the agent's logic flow and do not invoke arbitrary shell commands or access sensitive system files.
[DATA_EXPOSURE_EXFILTRATION]: The skill references local configuration files like brand-config.json and SOUL.md. These are specific to the skill's domain (branding and messaging) and no exfiltration patterns or external network requests were detected.
[EXTERNAL_DOWNLOADS]: No external URLs or remote dependencies are present in the provided content.
[PROMPT_INJECTION]: The skill does not contain instructions attempting to override safety guidelines or reveal system prompts.

evp-kickoff