The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references an installation script for the belt CLI from https://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md. This is a documentation reference to a well-known service's repository and is considered safe.
[COMMAND_EXECUTION]: The skill provides numerous examples of using the belt CLI (e.g., belt app run) and shell loops (for i in {1..5}; do ... done) for logo generation tasks. These are instructional examples for the user or agent to follow and do not execute dangerous operations.
[INDIRECT_PROMPT_INJECTION]: While the skill processes user-provided prompts for image generation (e.g., descriptions of logos), the output is a raster image, which cannot influence the agent's logic or downstream text-based execution pipelines. The surface for injection is negligible.
[REMOTE_CODE_EXECUTION]: The skill mentions npx skills add belt-sh/cli and other npx skills add commands. These are standard package management commands for installing skills from a well-known service and do not represent a security risk in this context.

logo-design-guide