gemini-deep-research

Automate Gemini Deep Research workflows and extract reports as Markdown.

Updated Feb 19, 2026
One-click install
npx skills add https://github.com/kzarzycki/agent-skills --skill gemini-deep-research-kzarzycki
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: gemini-deep-research
Source: https://github.com/kzarzycki/agent-skills/tree/main/research/skills/gemini-deep-research
Command: npx skills add https://github.com/kzarzycki/agent-skills --skill gemini-deep-research-kzarzycki

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

It solves the problem of performing repeatable, platform-native Gemini Deep Research runs (with the right UI configuration) and getting the resulting report reliably extracted as Markdown.

Core Features & Use Cases

  • Standalone Deep Research Runner: Spawns a browser-researcher agent to submit your research prompt on Gemini and capture the final report output.
  • Automated UI Flow: Drives Gemini’s Deep Research workflow end-to-end (login check, start, plan handling, approval, completion, and extraction).
  • Structured Report Extraction: Extracts the Gemini report panel content and converts Gemini-specific HTML into cleaned Markdown.
  • Use Case: When you need a cited research report from Gemini with consistent interaction steps, you can run it without manually clicking through Tools → Deep Research → Pro, then copy the finalized results.

Quick Start

Ask for “run deep research on <your topic> using Gemini” so the skill activates Gemini Deep Research, submits your prompt, and returns the extracted Markdown report.

Frequently Asked Questions about gemini-deep-research

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

FAQPage Schema
How do I automate Gemini Deep Research to extract a Markdown report?

Automating Gemini Deep Research requires driving the platform-native UI end-to-end using browser automation. This workflow submits your prompt, handles plan review and approval, and extracts the finalized research panel content into cleaned Markdown.

What is the best way to run platform-scoped Deep Research on Gemini without manual clicking?

The best way to run Gemini Deep Research without manual clicks is using a persistent Chromium profile via Playwright-CLI. This detects login sessions, automates multi-step plan handling, and captures the deterministic report output directly from the DOM.

Do I need browser automation to get a structured Markdown output from Gemini Deep Research?

Yes, browser automation is required to interact with Gemini's built-in Deep Research UI and capture the dynamic report panel. A Playwright-CLI driven browser-researcher agent converts the Gemini-specific HTML structure into structured Markdown.

How does the automated plan handling work during a Gemini Deep Research run?

Automated plan handling drives the multi-step UI workflow by submitting the research prompt, detecting the generated research plan, programmatically approving it, and waiting for completion before extracting the final cited report.

Can I use Playwright-CLI to capture research reports from Gemini directly?

Yes, Playwright-CLI can capture Gemini research reports by driving a persistent Chromium profile. It executes the full Deep Research workflow, performs login and session detection, and uses a DOM-based HTML-to-Markdown converter for extraction.

Why does my automated web research extraction fail to capture the full Gemini report?

Extraction fails if the automation does not wait for the full Deep Research completion or lacks proper session detection. Reliable extraction requires driving the end-to-end UI flow and applying a Gemini-specific HTML-to-Markdown converter to the finalized DOM.