The Agent Skills Directory

[SAFE]: The skill consists entirely of markdown documentation and instructional guidelines. No executable code, scripts, or configuration files that could perform actions are included.
[DATA_EXPOSURE]: While the skill mentions using an external API (Higgsfield), it does not contain or request hardcoded credentials. It references local project files (e.g., brand-pack.md) for context, which is standard behavior for agent skills.
[INDIRECT_PROMPT_INJECTION]: The skill describes workflows involving user-provided inputs (descriptions, references). While this represents a surface for indirect injection, there are no dangerous capabilities (like command execution or network exfiltration) defined within the provided skill content.

mcp-higgsfield