The Agent Skills Directory

[SAFE]: The skill performs its stated function of content repurposing using established tools and services without any detected malicious patterns.
[COMMAND_EXECUTION]: The instructions utilize the belt CLI tool to perform actions like image generation and social media posting, which is controlled via the allowed-tools configuration in SKILL.md.
[EXTERNAL_DOWNLOADS]: The skill references the installation of the belt CLI and additional skills from the inference-sh ecosystem to support its functionality. These resources are consistent with the vendor's own infrastructure.
[PROMPT_INJECTION]: The skill processes user-supplied content (such as blog posts or podcast transcripts) to generate inputs for subsequent tool calls. This represents an attack surface for indirect prompt injection, which is a common characteristic of content-processing skills.
Ingestion points: User-provided content intended for repurposing in SKILL.md.
Boundary markers: The skill uses structured JSON objects for tool inputs, which helps separate data from commands.
Capability inventory: Through the belt CLI, the skill can access apps for image generation, text-to-speech, and social media posting as seen in the recipes in SKILL.md.
Sanitization: The instructions do not specify explicit sanitization steps for the input content before it is used in tool prompts.

content-repurposing