The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to download and install external code and configuration from non-trusted sources.
Fetches installation instructions/scripts from https://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md.
Downloads a Node.js package using npx skills add belt-sh/cli.
[REMOTE_CODE_EXECUTION]: The skill uses npx to execute external code from the belt-sh/cli package, which is a high-risk pattern as it downloads and runs code from an external registry without explicit version pinning or integrity verification in the instructions.
[COMMAND_EXECUTION]: The skill is configured to execute shell commands using the belt CLI tool via the Bash tool wrapper. This provides a direct path for executing operations on the host system, scoped by the belt * pattern.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it ingests untrusted external data and incorporates it into command execution.
Ingestion points: The skill accepts user-provided values for prompt, image_url, audio_url, and video_url (SKILL.md).
Boundary markers: No boundary markers or delimiters are used to separate untrusted data from instructions.
Capability inventory: The skill can execute shell commands via Bash(belt *) (SKILL.md).
Sanitization: There is no evidence of input validation, escaping, or sanitization of the external content before it is passed to the shell environment.

ai-video-generation