The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute the infsh CLI. This is the intended functionality and is restricted to the specific vendor utility, following the principle of least privilege.
[EXTERNAL_DOWNLOADS]: The documentation mentions installing additional components via npx. These resources originate from the vendor's own verified ecosystem (inference-sh) and are standard installation procedures for this platform.
[PROMPT_INJECTION]: The skill interfaces with LLMs, which inherently involves processing untrusted user input via the --input flag. This represents a known surface for indirect prompt injection, but it is the primary purpose of the skill and is managed by the model provider's safety filters.

llm-models