design-system-gtm
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides a local Python tool (
scripts/check-output.py) to verify that the design system outputs contain all required sections. This script operates locally and only reads the contents of the generated files to match header strings. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest brand identity and voice data from users, which represents a potential attack surface.
- Ingestion Points: User-provided brand context is integrated into the output through the step-by-step process and the provided
templates/output-template.md. - Boundary Markers: No explicit delimiters or guardrail instructions are used to separate the user-supplied data from the agent's operational prompts.
- Capability Inventory: The agent can execute a local script (
scripts/check-output.py) to read files from the filesystem. - Sanitization: The skill does not implement sanitization or validation logic for the brand guidelines it processes.
Audit Metadata