desktop-workflow-generator
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill's primary function is to assist in QA documentation and testing through codebase exploration and interactive browser sessions.
- [CREDENTIALS_UNSAFE]: The skill manages authentication states using Playwright's storageState format (cookies and local storage). These are handled locally within the project's
.playwright/profiles/directory and are not hardcoded or exfiltrated. - [INDIRECT_PROMPT_INJECTION]: The skill reads external data from the application's source code and live UI. While this introduces a potential injection surface, the skill implements a multi-phase co-authoring process where the human user must review and approve all findings, screenshots, and workflow steps, significantly mitigating the risk of automated obedience to malicious embedded instructions.
- [DYNAMIC_EXECUTION]: The instructions include JavaScript templates for manipulating browser state (e.g.,
page.evaluate). These are standard automation patterns and are not used to execute untrusted code from remote sources.
Audit Metadata