fixing-motion-performance
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The instructions contain standard constraints for an AI agent's task, such as applying specific rules to UI animation work and limiting the scope of reviews. No attempts to override safety filters, bypass guidelines, or extract system prompts were detected.
- [DATA_EXPOSURE]: No hardcoded credentials, sensitive file paths, or network operations were found in the skill. The code examples use standard, benign browser APIs for performance measurement.
- [REMOTE_CODE_EXECUTION]: There are no commands or instructions that download and execute external scripts. The skill explicitly advises against migrating libraries unless requested, limiting the risk of introducing new dependencies.
- [DYNAMIC_EXECUTION]: The JavaScript code snippets provided are illustrative examples for the agent to suggest as fixes for the user. They are not intended for execution within the agent's own environment.
- [INDIRECT_PROMPT_INJECTION]: While the skill takes user-provided files as input for review, its capabilities are restricted to generating textual analysis and code suggestions. It does not have write access or the ability to execute the contents of the files being reviewed.
Audit Metadata