stelow-workflow-ux-critique
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from live URLs, source code directories, and image files, creating a surface for indirect prompt injection.
- Ingestion points: Processes external data from a
URL(Live Site Audit mode), asource directory(Codebase Audit mode), or ascreenshotimage (Screenshot Audit mode). - Boundary markers: The instructions do not explicitly define boundary markers or delimiters to isolate untrusted content from the agent's instructions when passing data to the reviewer subagent.
- Capability inventory: The skill leverages
agent_browserfor live site interaction,subagentsfor analysis and report generation, and shell commands (find,git diff) for file system discovery inSKILL.md. - Sanitization: There is no evidence of sanitization, filtering, or validation of external content before it is processed by the analysis subagent.
Audit Metadata