The Agent Skills Directory

REMOTE_CODE_EXECUTION (CRITICAL): The skill documentation explicitly recommends installing the infsh CLI using curl -fsSL https://cli.inference.sh | sh. This pattern is a major security risk as it executes unverified code from a non-trusted domain directly in the user's terminal.
EXTERNAL_DOWNLOADS (HIGH): The skill relies on binaries and scripts hosted at inference.sh and dist.inference.sh. These domains are not part of the trusted source list, meaning the integrity and safety of the downloaded files cannot be guaranteed.
COMMAND_EXECUTION (HIGH): The skill defines allowed-tools: Bash(infsh *), which grants the agent broad permission to execute any subcommand of the infsh utility, potentially leading to unintended local system operations or data exposure.
DYNAMIC_EXECUTION (MEDIUM): The core functionality of the skill involves taking arbitrary React/TSX code as a string input and sending it to a remote service for rendering. This pattern of 'string-to-code' execution increases the surface for indirect prompt injection, where malicious instructions could be embedded in the code passed to the renderer.

remotion-render