deep-research

Plan, search, decide, and synthesize multi-hop research over an rlat knowledge corpus.

16|1|Updated Apr 9, 2026
One-click install
npx skills add https://github.com/tenfingerseddy/resonance-lattice --skill deep-research-tenfingerseddy
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: deep-research
Source: https://github.com/tenfingerseddy/resonance-lattice/tree/main/.claude/skills/deep-research
Command: npx skills add https://github.com/tenfingerseddy/resonance-lattice --skill deep-research-tenfingerseddy

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Interactive multi-hop research over an rlat knowledge model, run by Claude in this session, to surface evidence, reason about sources, and synthesize final answers without switching contexts.

Core Features & Use Cases

  • Plan → retrieve → refine → retrieve → synthesize loop that guides the user through questions requiring cross-source analysis.
  • Trigger for memory recall across past sessions and for cross-workstream reasoning.
  • Inline citations and source-traceability, with evidence hierarchy across files.

Quick Start

Ask a multi-source question to start the deep-research loop and guide it through plan, search, decide, and synth steps.

Frequently Asked Questions about deep-research

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

FAQPage Schema
How do I synthesize evidence from multiple sources for cross-workstream research?

To synthesize cross-workstream research, use the multi-hop loop to plan, search, decide, and synth steps. This surfaces evidence from multiple sources, traces reasoning across files, and identifies contradictions for robust conclusions.

What is multi-hop research and how does it work with knowledge corpora?

Multi-hop research over a knowledge corpus involves a four-hop loop: plan, search, decide, and synth. It traces reasoning across files to surface evidence and gaps, ensuring robust conclusions without switching contexts.

How do I trace reasoning across documents to find contradictions or gaps?

Trace reasoning across documents by triggering the deep-research loop within a Claude session. It enforces a plan-search-decide-synth cycle with inline citations and strict stopping conditions to surface contradictions and gaps.

Can I use Claude to retrieve and cross-reference evidence from past sessions?

Yes, you can trigger memory recall across past sessions for cross-workstream reasoning. The interactive loop retrieves and cross-references evidence with inline citations and source-traceability across your knowledge model.

How to prevent over-searching when running multi-source analysis queries?

Prevent over-searching using strict stopping conditions enforced during the decide phase of the research loop. These guards halt retrieval once sufficient evidence is surfaced, avoiding over-search and incomplete results.

Does deep-research require external dependencies to synthesize knowledge models?

No, deep-research requires no external dependencies to synthesize knowledge models. It runs entirely within a Claude session over an rlat knowledge corpus to surface evidence and generate final answers interactively.