The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the belt CLI tool for account authentication (belt login) and to run various AI models (belt app run) for image generation and data processing.
[EXTERNAL_DOWNLOADS]: The skill references the installation of the @belt-sh/cli package and links to installation instructions hosted on a remote GitHub repository (inference-sh/skills). These resources are associated with the skill's authoring vendor.
[INDIRECT_PROMPT_INJECTION]: The skill uses web search tools (tavily/search-assistant, exa/search) to research competitor launches and community sentiment. This creates a surface where the agent could ingest and potentially follow malicious instructions embedded in external web content.
Ingestion points: Data retrieved from external web search results via tavily and exa applications.
Boundary markers: The prompts for search results do not include specific delimiters or instructions to ignore embedded commands.
Capability inventory: The skill has access to command execution via the belt CLI and can perform network operations via its applications.
Sanitization: There are no explicit sanitization or filtering steps defined for the processed search results.

product-hunt-launch