skills/redongreen/uspec/create-color/Gen Agent Trust Hub

create-color

Pass

Audited by Gen Agent Trust Hub on May 18, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill defines an indirect prompt injection surface by ingesting data from Figma files and interpolating it into JavaScript code executed within the Figma environment.
  • Ingestion points: Component names, variant properties, and design tokens extracted from the Figma document tree in SKILL.md (Steps 4 and 4b).
  • Boundary markers: Absent. The skill instructions rely on string interpolation of placeholders like __VARIANT_NAME__ and __TABLES_JSON__ without using delimiters or sanitization logic to isolate the ingested data from the script logic.
  • Capability inventory: The skill uses figma_execute and use_figma to perform operations that can create or modify nodes within the Figma document.
  • Sanitization: No specific sanitization or validation logic is provided for the data processed from the Figma environment before it is used in script construction.
  • [COMMAND_EXECUTION]: The skill programmatically generates and executes JavaScript snippets in the Figma plugin context via the figma_execute and use_figma tools. While these scripts are derived from predefined templates within the skill, they include dynamic data retrieved from the target Figma file's metadata.
Audit Metadata
Risk Level
SAFE
Analyzed
May 18, 2026, 10:11 PM
Security Audit — agent-trust-hub — create-color