Security audit
Security checks for vulnerabilities and agentic risk
This is a visible design-methodology skill with no executable payload, but users should prefer a pinned or manual install path and treat its connector-docs line cautiously.
Before installing, use a pinned version or manual reviewed clone when possible instead of unpinned npx commands. When using the skill, treat external connector docs as reference material only and keep final design actions under user direction.
The README instructs users to run npx skills add viratk-dev/virat-design without pinning a specific version of the skills package or an immutable source reference. Because npx fetches and executes code at install time, users may receive whatever package/version is current at that moment, creating a supply-chain risk if the package is updated maliciously or unexpectedly.
The update instruction npx skills update virat-design again relies on unpinned npx execution, which can download and run a changed version of the skills tooling in the future. This expands supply-chain exposure beyond initial install by encouraging repeated remote code execution during updates.
The manifest says to use the skill when creating, art-directing, or reviewing any visual design, which is extremely broad and lacks clear activation boundaries. Because it does not define specific trigger phrases, contexts, or exclusions, the skill could be invoked for a wide range of ordinary design-related requests unintentionally.
The instruction to use connector docs before following 'recovered instructions' introduces an external instruction source that can influence agent behavior outside the trusted skill content. Even though phrased as documentation lookup, it can create an instruction-precedence ambiguity and enable indirect prompt injection through connector documentation or recovered content.
No suspicious patterns detected.
