social-media-carousel

Fail

Audited by Gen Agent Trust Hub on Feb 19, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION] (HIGH): The skill instructs the user to execute curl -fsSL https://cli.inference.sh | sh. This is a classic piped-to-shell attack vector from an untrusted domain.
  • [EXTERNAL_DOWNLOADS] (HIGH): The installation process downloads an unverified binary from dist.inference.sh.
  • [COMMAND_EXECUTION] (HIGH): The skill uses npx to install additional remote components (inference-sh/skills), which facilitates arbitrary code execution from external sources.
  • [PROMPT_INJECTION] (LOW): The skill exhibits an indirect prompt injection surface. It ingests user-provided text/HTML and interpolates it directly into CLI commands without sanitization or boundary markers. Evidence Chain: 1. Ingestion points: Carousel text within HTML templates in SKILL.md. 2. Boundary markers: Absent. 3. Capability inventory: Access to Bash and the infsh CLI tool. 4. Sanitization: Absent.
Recommendations
  • HIGH: Downloads and executes remote code from: https://cli.inference.sh - DO NOT USE without thorough review
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 19, 2026, 08:30 PM