The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references external installation instructions and configuration files hosted on the inference-sh GitHub repository (raw.githubusercontent.com/inference-sh/).\n- [COMMAND_EXECUTION]: The skill executes shell commands using the belt CLI tool to perform searches, log in to services, and run external applications via the Bash tool.\n- [DATA_EXFILTRATION]: User queries and extracted content are transmitted to external services (Tavily, Exa, and OpenRouter) via the inference.sh infrastructure as part of the skill's primary functionality.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external websites, creating a surface for potential injection attacks from untrusted content.\n
Ingestion points: Web content retrieved through tavily/extract and exa/extract apps (SKILL.md).\n
Boundary markers: The skill recommends using tags like <search-results> or <content> in prompts, which provide minimal separation between data and instructions.\n
Capability inventory: The skill possesses file-writing capabilities and shell command execution (belt) across its workflows.\n
Sanitization: No sanitization, filtering, or validation of the fetched web content is performed before passing it to downstream language models.

web-search