By id: zai-deep-research name: Z.ai Deep Research Protocol description: Multi-source research protocol with verification, synthesis, and citation requirements. type: research scope: general priority: 90 contexts: [general, agency-research]
Z.ai Deep Research Protocol
When This Skill Activates
Activate for any research, investigation, fact-finding, market analysis, or "find out about X" task.
Research Stack (in order)
- z.ai native web search — real-time web results synthesized by GLM-5-Air with recency filter
- Brave Search API — independent search index for cross-verification
- Page Reader — auto-fetch and extract full text from top result URLs
- GLM-5-Thinking — activate chain-of-thought reasoning to synthesize findings
Research Protocol
Step 1 — Scope Definition
Before searching, define:
- Core question (what exactly needs answering)
- Time sensitivity (is recency important?)
- Evidence type needed (statistics, opinions, facts, comparisons)
Step 2 — Multi-Source Search
Run searches across all available tools:
- z.ai search: main query + sub-questions
- Brave: verify claims from z.ai results
- Page reader: deep-read top 2-3 URLs for full article context
Step 3 — Synthesis & Output Structure
Deliver findings as: ```
Executive Summary
[2-3 sentence bottom line]
Key Findings
- Finding 1 [source: url]
- Finding 2 [source: url]
- Finding 3 [source: url]
Supporting Details
[Detailed breakdown with citations]
Confidence Level
[HIGH/MEDIUM/LOW] — based on source quality and agreement across sources
Sources
Critical Rules
- ALWAYS cite sources with actual URLs
- ALWAYS distinguish fact from inference
- NEVER present search results as your own knowledge
- Flag when information is older than 30 days for time-sensitive topics
- Cross-verify controversial claims with 2+ independent sources
Thinking Mode
For complex multi-step research questions: use GLM-5-Thinking (model: glm-5-thinking) to reason through conflicting evidence and draw conclusions.