motion
Pass
Audited by Gen Agent Trust Hub on Nov 21, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The
tools/tts.mjsutility script is configured to automatically download and install the@speech-sdk/corepackage (version 0.38.0) from the npm registry if it is not already present. This occurs at runtime during the voiceover generation process. - [COMMAND_EXECUTION]: The skill uses the Node.js
child_process.spawnSyncmethod to execute system commands, specifically for runningnpm installto manage its text-to-speech dependencies and to invoke its own CLI scripts. - [DATA_EXFILTRATION]: The skill's
uploadcapability permits the agent to read files from absolute system paths and store them as binary blobs in the artifact's database. This provides a mechanism for accessing and potentially exfiltrating sensitive local files if the agent is directed to upload them. - [INDIRECT_PROMPT_INJECTION]: The skill architecture involves processing data from various external sources, including user-submitted comments, design systems, and manual edits. To mitigate potential injection attacks, the skill contains explicit, repetitive instructions across multiple files (e.g.,
SKILL.md,reference/design-systems.md,reference/by-hand.md) to treat all such input strictly as data and never as commands or instructions. - [DYNAMIC_EXECUTION]: The skill's primary function is the generation and execution of JavaScript code to render animation scenes. While this is inherently dynamic, the documentation describes a specialized runner that enforces constraints such as blocking network access and freezing timers during the rendering process.
Audit Metadata