The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references documentation and installation scripts from the inference-sh GitHub repository and official domain. These are standard resources for the described services and are documented neutrally as legitimate external dependencies.
[COMMAND_EXECUTION]: The skill utilizes the 'belt' CLI tool for interacting with external search and extraction APIs. Access is restricted via the 'allowed-tools' parameter in the YAML frontmatter, which scopes the agent's execution environment to this specific tool according to best practices.
[PROMPT_INJECTION]: The skill defines workflows that ingest external web content for processing by LLMs. While this creates an indirect prompt injection surface inherent to search-based agents, the skill includes examples of using boundary markers (e.g., , ) to help the model distinguish between instructions and ingested data.

web-search