forge-design
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill's logic is strictly confined to design token generation and code enforcement. No network access, sensitive file reads, or elevated permissions are requested.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for indirect prompt injection by reading a project-level configuration file (
.forge.json) and using its contents to modify source code. - Ingestion points: Reads
.forge.json(identified in SKILL.md). - Boundary markers: None specified.
- Capability inventory: Writes and modifies frontend code files (CSS, Tailwind, SCSS, JSON) and performs automated corrections.
- Sanitization: None specified in the instructions. This surface is standard for project-aware development tools and is assessed as safe given the specific domain of design tokens.
Audit Metadata