figma-to-code
Pass
Audited by Gen Agent Trust Hub on Apr 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a shell command (
ls) to check for the presence of recommended companion skills in various local directories (e.g.,~/.claude/skills/,~/.agent/skills/). This is used for environment verification and dependency management within the skill framework. - [EXTERNAL_DOWNLOADS]: The documentation and the companion check logic recommend the installation of additional skills from the
AbsolutelySkilledrepository using thenpx skills addcommand. - [PROMPT_INJECTION]: The skill includes a specific behavioral instruction for the agent to always start its first response with the 🧢 emoji. This is a benign branding instruction and does not attempt to bypass safety guidelines.
- [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface as it is designed to process untrusted data from Figma mockups, screenshots, and design specifications.
- Ingestion points: UI designs from Figma mockups, screenshots, or design specs (SKILL.md).
- Boundary markers: None provided in the instructions.
- Capability inventory: Filesystem listing via
ls(SKILL.md), recommendation of external code installation vianpx(README.md, SKILL.md). - Sanitization: No specific sanitization or validation logic is defined for the interpreted design data.
Audit Metadata