youtube-thumbnail-design

Fail

Audited by Gen Agent Trust Hub on Feb 18, 2026

Risk Level: CRITICALREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • Remote Code Execution (CRITICAL): The skill instructs the user/agent to execute curl -fsSL https://cli.inference.sh | sh. This pattern is highly dangerous as it bypasses package management safety and executes unverified remote code with shell privileges. The domain inference.sh is not a trusted source.
  • External Downloads (HIGH): The skill references and installs external components using npx skills add inference-sh/skills@.... These are untrusted dependencies that could contain malicious logic and are executed via Node.js.
  • Command Execution (MEDIUM): The skill's metadata grants broad access to the Bash tool (allowed-tools: Bash(infsh *)), which, when combined with the remote installation of the infsh CLI, creates a large attack surface for command injection.
  • Indirect Prompt Injection (LOW): The skill processes user intentions to generate image prompts without boundary markers or sanitization, creating a surface for indirect prompt injection.
  • Ingestion points: Prompt templates within SKILL.md (e.g., in the A/B testing and pattern sections).
  • Boundary markers: None identified.
  • Capability inventory: Execution of remote AI models via Bash(infsh *).
  • Sanitization: No escaping or validation of user-provided content before interpolation into model prompts.
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
CRITICAL
Analyzed
Feb 18, 2026, 01:59 AM
Security Audit — agent-trust-hub — youtube-thumbnail-design