social-media-carousel
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 command
curl -fsSL https://cli.inference.sh | shdownloads and executes code from a non-trusted third-party domain. This 'curl pipe bash' pattern is a critical vulnerability that permits arbitrary code execution and full system compromise. - EXTERNAL_DOWNLOADS (HIGH): The skill installs multiple unverified packages from the
inference-shorganization usingnpx. Since this organization is not on the list of trusted sources, these dependencies represent an unverified supply chain risk. - COMMAND_EXECUTION (MEDIUM): The skill generates and executes shell commands and loops to interact with the
infshCLI. The use of shell execution for dynamically generated templates increases the attack surface for potential exploits. - PROMPT_INJECTION (LOW): The skill is vulnerable to indirect prompt injection because it interpolates user-provided text into HTML slide templates without sanitization or boundary markers.
- Ingestion points: User-provided carousel content and layout parameters in
SKILL.md. - Boundary markers: Absent from HTML and bash templates.
- Capability inventory: Broad subprocess execution permissions via
Bash(infsh *). - Sanitization: No escaping or validation of external content is performed before interpolation.
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