Skills
skills/inf-sh/skills/social-media-carousel/Gen Agent Trust Hub

social-media-carousel

Fail

Audited by Gen Agent Trust Hub on Feb 18, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • REMOTE_CODE_EXECUTION (HIGH): The skill instructs the agent to run curl -fsSL https://cli.inference.sh | sh to install the infsh CLI. This piped remote execution pattern is dangerous because it executes unverified code from an untrusted domain directly in the shell context.
  • EXTERNAL_DOWNLOADS (MEDIUM): The skill uses npx skills add inference-sh/skills@... to pull in additional skill definitions. The inference-sh organization is not on the list of trusted GitHub repositories or organizations, increasing the risk of supply chain attacks.
  • COMMAND_EXECUTION (LOW): The skill relies on the infsh tool and Bash execution to perform its primary tasks. While functional for the carousel use case, this requires trusting the inference.sh platform's binary safety.
  • PROMPT_INJECTION (LOW): The skill has an indirect prompt injection surface when processing user-provided HTML for image generation. \n
  • Ingestion points: The html field inside the JSON --input flag for infsh commands in SKILL.md.\n
  • Boundary markers: None. There are no delimiters or instructions provided to the agent to treat the HTML as untrusted data.\n
  • Capability inventory: The skill has access to the Bash tool with broad execution capability for infsh apps.\n
  • Sanitization: None. The skill does not provide mechanisms to sanitize or escape HTML before processing.
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 18, 2026, 01:59 AM