| name | x-ipe-tool-web-search |
| description | Delegate focused web research to the current coding agent's web capability, then return structured findings and citations. Use when a skill needs industry best practices, standards, competitor patterns, or library/API guidance from the web. Triggers on requests like "web search", "research topic", "find best practices", "search the web". |
Web Search Tool
Purpose
AI Agents follow this skill to perform focused external research by:
- Determining whether a question is appropriate for web research
- Delegating search and retrieval to the current agent's web capability
- Synthesizing findings into reusable principles with citations
Important Notes
BLOCKING: Use this skill ONLY for general knowledge topics such as industry standards, best practices, regulations, competitor patterns, or public API/library guidance.
BLOCKING: Do NOT use this skill for project-internal questions about repository code, local files, workflow state, or unpublished requirements.
CRITICAL: Prefer the current coding agent's native web capability first. Examples include built-in web search, internet-enabled fetch tools, or dedicated browsing/fetch tools exposed by the agent runtime.
CRITICAL: Keep research bounded. Default to 3 useful sources maximum unless the caller explicitly needs more.
About
This tool skill standardizes how X-IPE skills ask an agent to perform public web research. It does not invent a separate crawler or search engine. Instead, it tells the executing agent to use whatever web capability the CLI/runtime already provides, then normalize the result into a compact, reusable structure.
Key Concepts:
- General Knowledge Request — A topic that can be answered from public web sources without needing project-private context
- Agent Web Capability — Any built-in web search, browsing, fetch, or retrieval ability available to the current CLI agent
