The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses belt app run commands throughout the documentation as examples for interacting with different AI models (Claude, FLUX, Veo). These commands are for demonstration purposes and are restricted to the belt CLI tool.
[EXTERNAL_DOWNLOADS]: The skill references installation scripts and additional skills from the vendor's GitHub repository (github.com/inference-sh/*) and official domain (inference.sh). These are well-known resources belonging to the skill author.
[INDIRECT_PROMPT_INJECTION]: While the skill demonstrates how to provide input to AI models, which is a potential attack surface, it does so in an educational context with hardcoded examples and does not automate the ingestion of untrusted external data in a way that would facilitate an attack.

prompt-engineering