better-typography
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [NO_CODE]: The skill is composed entirely of markdown documentation and JSON metadata. It does not include any executable scripts, binaries, or active code components, which eliminates the risk of direct malicious execution.
- [SAFE]: A thorough review of all instructions and metadata found no evidence of prompt injection, hidden commands, or data exfiltration attempts. All content is consistent with the stated purpose of improving product typography.
- [INDIRECT_PROMPT_INJECTION]: The skill instructions involve the agent reviewing external code and rendered content, which creates a potential attack surface.
- Ingestion points: Rendered page content and source code (SKILL.md).
- Boundary markers: Absent.
- Capability inventory: No high-risk tools or scripts are defined or called within the skill. The instructions are limited to visual analysis and styling fixes.
- Sanitization: Absent.
Audit Metadata