deep-research

Conduct multi-source research with triangulation, contradiction handling, and citation-backed synthesis.

55|41|Updated Aug 9, 2026
One-click install
npx skills add https://github.com/Rylaispirit/rylai-codex-hermes-skills --skill deep-research-rylaispirit
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: deep-research
Source: https://github.com/Rylaispirit/rylai-codex-hermes-skills/tree/main/skills/deep-research
Command: npx skills add https://github.com/Rylaispirit/rylai-codex-hermes-skills --skill deep-research-rylaispirit

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Answering complex questions that require multiple sources often leads to shallow summaries, unverified claims, or fabricated citations. This Skill enforces a systematic research workflow that produces source-backed, citation-ready answers with explicit handling of contradictions and evidence gaps. ## Core Features & Use Cases - Scoped Research Planning: Breaks a broad question into independent sub-questions with defined scope, date boundaries, audience, and preferred primary sources. - Evidence Triangulation: Builds an evidence table tracking claim, source, date, confidence, and contradictions, then explains disagreements instead of silently picking one source. - Citation-Ready Synthesis: Attaches citations to the specific claims they support and separates verified fact from inference, stale evidence, and unresolved gaps. - Use Case: Ask whether a market trend is real or hype. The Skill searches broadly, inspects primary sources, triangulates conflicting reports, and delivers a report where every claim carries a citation and confidence note. ## Quick Start Use the deep-research skill to investigate whether remote work adoption is still growing in 2024 and produce a cited, source-backed report.

Frequently Asked Questions about deep-research

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

FAQPage Schema
How do I do deep research with an AI agent?▼

Define the research question, scope, and date boundary first, then break it into independent sub-questions. Search broadly, inspect primary sources, build an evidence table with confidence notes, and synthesize the answer with citations attached to each claim.

How to handle contradictory sources in research?▼

Triangulate important claims across multiple independent sources and record disagreements in an evidence table. Explain the contradiction explicitly in the final report rather than silently selecting one source as correct.

Can AI research assistants invent citations?▼

Poorly constrained agents can fabricate citations, URLs, or statistics. This Skill explicitly forbids inventing citations, quotes, dates, or source conclusions, and requires verifying freshness of current topics at execution time.

When should I use parallel research agents?▼

Use parallel research agents only when the runtime supports delegation and the subtasks are genuinely independent. For tightly coupled questions, sequential investigation avoids duplicated effort and conflicting partial results.

What are the limitations of automated deep research?▼

Results depend on available search tools, network access, and source accessibility at runtime. Paywalled or offline sources may be unreachable, and the report must state unresolved gaps and stale evidence separately from verified facts.