video-ad-specs
Fail
Audited by Gen Agent Trust Hub on Feb 19, 2026
Risk Level: CRITICALREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION] (CRITICAL): The skill documentation includes the command
curl -fsSL https://cli.inference.sh | shin the Quick Start section ofSKILL.md. This is a high-risk pattern that fetches and executes an unverified script from the internet directly in the user's shell, allowing for complete system compromise. - [EXTERNAL_DOWNLOADS] (HIGH): The skill relies on binaries and scripts hosted at
cli.inference.shanddist.inference.sh, which are not included in the pre-approved trusted source list. - [COMMAND_EXECUTION] (MEDIUM): The skill utilizes a Bash tool to run the
infshCLI and usesnpx skills addto fetch further skills from external repositories, both of which introduce risks when associated with an untrusted installer. - [PROMPT_INJECTION] (LOW): Vulnerable to Indirect Prompt Injection (Category 8). The skill interpolates untrusted user-supplied prompts into
infshtool calls (seen inSKILL.mdexamples) without explicit sanitization, boundary markers, or warnings to the model to ignore embedded instructions.
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