The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires installing external packages and CLI tools from belt-sh and inference-sh via npx and references installation instructions hosted on a remote GitHub repository (https://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md).
[COMMAND_EXECUTION]: The skill utilizes the belt CLI tool within a Bash environment to perform network-based search operations and interact with remote APIs.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process content from external websites, which presents an attack surface where malicious instructions embedded in web pages could influence the agent's behavior.
Ingestion points: Data enters the agent's context through Tavily/Exa search results and content extracted from user-provided URLs (SKILL.md).
Boundary markers: Uses simple markdown-like tags such as <search-results> and <content> in example prompts, which provide minimal protection against adversarial content injection (SKILL.md).
Capability inventory: The skill uses the belt CLI to run various search and extraction apps, with results often piped into LLMs for summarization or analysis (SKILL.md).
Sanitization: There is no evidence of content sanitization, filtering, or escaping for the data retrieved from the web before it is interpolated into prompts (SKILL.md).

web-search