The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill documentation references an external installation guide hosted on GitHub (inference-sh/skills) and recommends the installation of the belt-sh/cli tool. It also includes code examples utilizing the inferencesh Python library.
[COMMAND_EXECUTION]: The skill instructions demonstrate how to use the belt CLI tool to perform image generation and manipulation tasks.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted user input in the form of text prompts and images, which are then passed to the Gemini model via the inference service. This represents a standard surface for indirect prompt injection.
Ingestion points: The prompt and images parameters within the JSON input for the belt app run command in SKILL.md.
Boundary markers: No explicit boundary markers or delimiters are used to separate user-provided content from instruction context.
Capability inventory: The skill is configured to use the Bash tool to execute belt CLI commands.
Sanitization: There is no evidence of input sanitization or validation before the prompt data is processed by the CLI tool.

nano-banana-2