deep-research-t1

Gather, validate, and synthesize external knowledge via graph-based expansion and web search protocols.

Updated Apr 8, 2026

Search Tags:#playbook #web-search #knowledge-graph #deep-research #source-tiering #high-stakes #cok

Ready to run

npx skills add https://github.com/rd162/skills --skill deep-research-t1

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: deep-research-t1
Source: https://github.com/rd162/skills/tree/main/deep-research-t1
Command: npx skills add https://github.com/rd162/skills --skill deep-research-t1

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Gathers, validates, and synthesizes external knowledge using CoK graph-based expansion until saturation.

Core Features & Use Cases

CoK graph expansion: iterative, triple-based knowledge expansion with saturation checks.
Δ1-Δ7 web search protocol: end-to-end search, fetch, and synthesis pipeline.
High-stakes handling: forward-consequence CoK patterns and explicit safety/disclaimer generation.
Sub-agent fan-out: parallel research on multiple subjects with isolated contexts.

Quick Start

Provide a topic to research and an optional output mode; the system will run the Δ1-Δ7 protocol and return a structured playbook.

Frequently Asked Questions about deep-research-t1

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I validate and synthesize external knowledge from web searches?▼

To validate and synthesize external knowledge, apply a structured web search protocol with domain-specific source tiering (T1-T4) and iterative graph-based expansion until saturation is reached. This ensures gathered information is verified before synthesis.

What is a CoK graph-based expansion for deep research?▼

CoK graph-based expansion is an iterative, triple-based knowledge expansion method that performs saturation checks. It continuously gathers and validates external knowledge until no new relevant information is found, creating a comprehensive research playbook.

How do I handle contradictions in high-stakes research synthesis?▼

Handling contradictions in high-stakes research requires applying forward-consequence CoK patterns and generating explicit safety disclosures. The system identifies conflicting information during synthesis and flags it with appropriate disclaimers.

Can I research multiple subjects in parallel using sub-agent fan-out?▼

Yes, you can research multiple subjects in parallel using sub-agent fan-out, which creates isolated contexts for each subject. This allows simultaneous research streams without cross-contamination of information.

What's the best way to structure temporal-aware web search queries?▼

The best way to structure temporal-aware web search queries is through a Δ1-Δ7 search protocol that integrates time-sensitive parameters with domain-specific source tiering (T1-T4), ensuring results reflect the most current and relevant information available.

Do I need to provide a specific output mode to generate a research playbook?▼

No, providing an output mode is optional when generating a research playbook. You need to supply a research topic, and the system will automatically run the Δ1-Δ7 protocol and return structured results, with output mode being a customizable enhancement.