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

create-motion

Warn

Audited by Gen Agent Trust Hub on May 18, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses figma_execute and use_figma tools to run dynamically generated JavaScript code within the Figma environment. This code is constructed using placeholders filled at runtime with user-provided data and configuration values.
  • Evidence: Steps 8, 9, 10, 11, and 13 contain JavaScript blocks for importing components, manipulating text, and rendering timeline layers.
  • [PROMPT_INJECTION]: The skill features a surface for indirect prompt injection by ingesting untrusted JSON data from the clipboard or file references to drive its rendering logic.
  • Ingestion points: Clipboard JSON or file references (e.g., @motion-ex.json) read in Step 4.
  • Boundary markers: The skill instructions do not specify any markers or delimiters to isolate the ingested data from the agent's instructions.
  • Capability inventory: The agent has the ability to execute code in Figma, navigate to URLs, and take screenshots.
  • Sanitization: While Step 5 mentions JSON schema validation, it does not describe measures to sanitize content against malicious instructions.
  • [EXTERNAL_DOWNLOADS]: The skill fetches external assets by importing Figma components using keys stored in a local configuration file.
  • Evidence: Step 8 uses figma.importComponentByKeyAsync(TEMPLATE_KEY) where TEMPLATE_KEY is read from uspecs.config.json.
Audit Metadata
Risk Level
MEDIUM
Analyzed
May 18, 2026, 10:10 PM
Security Audit — agent-trust-hub — create-motion