ai-marketing-videos
Fail
Audited by Gen Agent Trust Hub on Feb 18, 2026
Risk Level: CRITICALREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- Remote Code Execution (CRITICAL): The skill documentation explicitly instructs the user to run
curl -fsSL https://cli.inference.sh | sh. This is a high-risk pattern that executes an unverified remote script with the user's shell privileges. - External Downloads (HIGH): The skill references and encourages adding multiple external skills using
npx skills add inference-sh/skills@.... This creates a chain of unverified dependencies from an untrusted source. - Command Execution (MEDIUM): The skill utilizes a powerful bash tool
infshwith broad permissions (Bash(infsh *)), allowing for arbitrary subcommand execution including login and resource management. - Indirect Prompt Injection (LOW): The skill processes user-controlled prompts by interpolating them directly into JSON payloads for AI models.
- Ingestion points: The
--inputJSON strings in multipleinfsh app runcommands. - Boundary markers: None detected; user input is directly concatenated into the command string.
- Capability inventory: Full bash execution via the
infshutility. - Sanitization: No evidence of sanitization or escaping for user-provided prompt strings.
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