The Agent Skills Directory

[SAFE]: The skill consists of natural language prompts focused on organizational and summarization tasks within a design tool context. No executable code, shell scripts, or unauthorized network requests are present.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from Figma design-file context (ingestion point: SKILL.md), such as frames, components, and comments. No explicit boundary markers are defined to isolate this untrusted content. However, the skill's capabilities are limited to report generation and scoped edits within the Figma environment, with no identified access to shell execution, network operations, or sensitive file system paths (capability inventory: Figma reporting/editing). No sanitization methods were mentioned in the instructions. The overall risk is negligible due to the restricted environment.

handoff-summary