tailwind-v4-shadcn
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOWSAFE
Full Analysis
- [Indirect Prompt Injection] (LOW): The skill is designed to trigger based on user-provided strings, including specific error messages and configuration snippets (e.g.,
tw-animate-css,@plugin directive). While these are intended as triggers for helpful automation, they represent an ingestion point where specifically crafted user input could influence the agent's behavior. - Ingestion points: README.md (Primary, Secondary, and Error-Based Triggers).
- Boundary markers: Absent in the analyzed README.md; relies on the agent's internal keyword matching.
- Capability inventory: The README.md describes a workflow for copying templates and following setup instructions in a secondary
SKILL.mdfile (not provided). The README itself does not contain executable scripts or network calls. - Sanitization: None explicitly mentioned in the README metadata or instructions.
- [Unverifiable Dependencies] (LOW): The skill lists specific versions for npm packages (e.g.,
tailwindcss@^4.1.14,vite@^7.0.0). These are standard, well-known libraries from the public npm registry. No suspicious or unknown packages were detected. - [Metadata Poisoning] (INFO): The metadata includes a link to an external 'Production Tested' site (
wordpress-auditor.webfonts.workers.dev). While not malicious in the context of the README, users should always be cautious of external links in metadata.
Audit Metadata