frontend-testing-debugging
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- Command Execution: The skill utilizes shell commands through package managers (e.g.,
pnpm exec playwright) to perform automated testing and capture screenshots. This capability allows the agent to interface with local development tools to verify the state of the application. - Dynamic Execution: To perform deeper debugging, the skill may generate and execute temporary Playwright scripts. This runtime script generation is a primary feature for validating complex UI interactions that are not covered by static tests.
- Indirect Prompt Injection Surface: The skill ingests data from external sources, including DOM snapshots and browser console logs. Because the agent interprets this data to generate QA reports, there is a potential for instructions embedded within the tested web application to influence the agent's analysis.
- File System Access: The workflow involves writing temporary artifacts, such as screenshots and scripts, to the local file system (e.g.,
/tmp/). These operations are scoped to stay outside of the committed source code to prevent repository pollution.
Audit Metadata