The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill directs users to install external CLI tools and supplementary skills using npx skills add belt-sh/cli and various repositories under the inference-sh/skills namespace.
[COMMAND_EXECUTION]: The skill relies on shell command execution via the belt utility, including commands for session authentication (belt login) and executing remote AI models (belt app run).
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied assets and text, which could contain instructions intended to influence the agent's behavior.
Ingestion points: The skill ingests external image URLs, audio URLs, and text-based scripts provided as arguments to the belt CLI tool in SKILL.md.
Boundary markers: There are no explicit delimiters or safety instructions provided to the agent to distinguish between the production instructions and the content of the user-provided data.
Capability inventory: The skill possesses the capability to execute shell commands and interact with remote APIs using the belt tool.
Sanitization: The skill does not implement or describe any validation or sanitization mechanisms for the input URLs or voice scripts.

talking-head-production