search-analyst-pair

What This Skill Does

The search-analyst-pair skill is a specialized architectural workflow for OpenClaw designed to transform ambiguous research requests into highly structured, actionable briefings. By utilizing a deterministic three-hop chain, it enforces a separation of concerns that ensures information quality. The workflow initiates via the /hunt trigger, which activates the search agent to perform deep data collection using DeepSeek. Once the information is gathered, it is passed to the analyst agent, powered by Gemini, which performs critical evaluation, risk assessment, and key point extraction. Finally, the main agent (Kimi) aggregates the findings and analysis to formulate a concise, professional conclusion. This multi-model orchestration ensures that facts remain verifiable, analysis remains critical, and the final output is synthesize for immediate decision-making. By delegating specific tasks to specialized subagents, the skill eliminates the variability often found in ad-hoc agent responses, making it ideal for high-stakes professional research.

Installation

To install the skill, run the following command in your terminal within the OpenClaw environment:

clawhub install openclaw/skills/skills/baobaodawang-creater/search-analyst-pair

Ensure your OpenClaw configuration has tools.agentToAgent.enabled=true and that the subagents main, search, and analyst are correctly defined in your agent registry. Verification of agent access rights is required before the first execution.

Use Cases

This skill is perfect for complex research tasks such as:

• Technical Documentation Audits: Compare legacy system frameworks against modern alternatives with structured risk analysis.
• Market Intelligence: Aggregate weekly updates in niche industries and analyze the impact on current business workflows.
• Legal and Compliance Research: Gather regulatory updates and categorize them by implementation difficulty and operational risk.

Example Prompts

1. /hunt Review the latest updates to the React 19 documentation and provide a summary of breaking changes versus new performance optimizations.
2. /hunt Analyze the current state of open-source vector databases and compare the operational risks of using Qdrant versus Milvus for production.
3. /hunt Summarize recent industry reports on AI agent security standards and offer a checklist for enterprise deployment.

Tips & Limitations

• Tip: Use specific keywords in your /hunt prompt to help the search agent narrow its initial query range.
• Tip: If the process hangs, verify that your API quotas for the specific models (DeepSeek, Gemini, Kimi) are sufficient for multi-step agent-to-agent communication.
• Limitation: The skill relies on external API connectivity; if an upstream model provider experiences downtime, the entire chain will fail. Ensure you handle potential fallback logs provided by the agent gateway.