skills/cabinet-fe/ultra-ui/cat-kit-fe/Gen Agent Trust Hub

cat-kit-fe

Pass

Audited by Gen Agent Trust Hub on Apr 23, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [SAFE]: The skill contains no executable code or scripts, serving only as a reference for the @cat-kit/fe library. The documentation and examples reflect standard browser development practices.
  • [DATA_EXFILTRATION]: The documented API provides access to sensitive browser-side data, including system clipboard contents via clipboard.readText() and all stored cookies via cookie.getAll(). These are legitimate utilities for browser-based agents but involve access to privacy-sensitive data stores.
  • [PROMPT_INJECTION]: The skill defines functions that ingest untrusted data into the agent's context, such as readChunks for files and clipboard.readText for the clipboard. Ingestion points: file contents and clipboard text (SKILL.md, generated/file/read.d.ts, generated/web-api/clipboard.d.ts). Boundary markers: none present in the type definitions. Capability inventory: standard browser APIs including storage access and file downloads. Sanitization: no sanitization logic is specified in the provided interfaces. This represents a potential surface for indirect prompt injection if the agent processes untrusted content.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 23, 2026, 02:22 PM
Security Audit — agent-trust-hub — cat-kit-fe