web-search

The Web Search and Research Specialist is a high-precision autonomous skill designed for deep information gathering and critical synthesis. Targeted at analysts, developers, and researchers, this agent moves beyond basic web scraping by applying rigorous verification protocols and intellectual frameworks to every query. It serves as an expert assistant capable of navigating fragmented information environments to provide concise, evidence-based answers. Unlike generic search bots, this agent focuses on source credibility, using CRAAP criteria (Currency, Relevance, Authority, Accuracy, Purpose) to filter out noise, SEO-heavy content farms, and unverified secondary opinions, ensuring that the retrieved data is grounded in primary documentation, academic papers, and official announcements. Users can expect the agent to handle complex, multi-step search requests with logical reasoning, managing discrepancies between conflicting sources by transparently documenting disagreements. It excels at technical documentation retrieval, competitive intelligence, and market analysis, where accuracy and recent context are paramount. The agent actively incorporates current-year modifiers and site-specific search syntax to ensure the most relevant results are captured without the need for manual prompt engineering iteration.

• Performs targeted web searches using precise query formulation and Boolean operators.

• Evaluates source authority and relevance, prioritizing official documentation, research papers, and verified technical reports.

• Synthesizes findings into structured reports, summarizing long-form content into actionable takeaways with clear URL citations.

• Manages information conflicts by presenting multiple perspectives and documenting discrepancies.

• Utilizes technical context including framework names, exact error messages, and version numbers for troubleshooting.

• Flags outdated, unreliable, or speculative information to maintain high trust levels.

• Operates as a research-centric autonomous agent that can be scheduled to run unattended.

• Input requirements include clear research objectives, specific topics, or technical parameters; provide context such as versions or specific goals to improve result relevance.

• Expected output is a structured summary that leads with a direct answer, followed by detailed supporting evidence and source documentation.

• Practice site-specific searches using standard syntax (e.g., site:docs.python.org) when working within known ecosystems to maximize result quality.

• Ensure that all research queries include the current year or version context if searching for evolving technologies like Rust, AI agents, or specific API changes.

• Avoid requesting broad, non-specific topics without clear constraints; the agent performs best when provided with a defined scope or objective.