Skill Listing

Review the skill and install it into your account in one click.

Sign in
← Back to Skills

Deep Research

Comprehensive research using multi-source evidence, synthesis, and citation structure.

Free

Category

research

Provider

computer agents

Code Files

1

View on GitHub

AI review

Quality 82Trust 68Discovery 51

A practical, well-documented CLI skill for multi-source research with streaming progress and output options; it appears functionally solid but depends on an external Google Gemini API key and lacks explicit privacy/provenance and hosting verification details.

Deep Research is a useful, well-documented command-line research agent that streams progress, supports retries, and can save comprehensive reports with citations. The listing provides clear usage examples, options (timeout, output, verbose), and a single runnable script, making it straightforward to try. Note that it requires a Google Gemini API key (or GEMINI_API_KEY env var) and should be run in a trusted environment; reviewers should confirm citation provenance, cost/rate-limit implications, and any data-handling/privacy policies before production use.

Source full AI review

Strengths

  • Clear, practical CLI with examples and configurable options (output, timeout, verbose).
  • Streaming progress updates and automatic retry/polling behavior improve usability for long-running research.

Considerations

  • Relies on an external Google Gemini API key; users must manage credentials and be aware of costs and rate limits.
  • Listing lacks explicit details on citation provenance, privacy/data handling, and domain/source verification for the hosted code.

Why this ranks

Agent List ranks listings using quality, trust, traction, and freshness instead of follower count alone. Paid Computer Agents badges are identity signals only and do not raise discovery score.

  • It is surfacing as a hidden-gem candidate: high quality with less existing traction.
  • The quality review is solid enough to support discovery.
  • The listing is under additional review, which can affect discoverability.

Trust signals

AI review

Review pending.

Badge guide

Paid account badge

Automatic for active Computer Agents Individual, Team, and Enterprise subscriptions. It confirms account status only and does not increase discovery ranking.

Agent List reviewed

Agent List review

Granted after Agent List reviews the creator profile and marketplace presence. This is separate from the blue paid account badge.

name:
deep-research
description:
Conduct comprehensive multi-phase research on complex topics. Use when asked to research, investigate, analyze, or create detailed reports on topics requiring multiple sources, in-depth analysis, or comprehensive coverage.

Deep Research

Conduct comprehensive deep research on complex topics using Google's Deep Research agent with streaming progress updates.

When to Use

Use this skill when you need to:

  • Research complex topics requiring multiple sources
  • Generate comprehensive reports with citations
  • Analyze and synthesize information from various angles
  • Investigate technical concepts, industry trends, or detailed questions

Usage

Run the deep research script with your topic:

terminal
Loading...

Optional: Save to a specific output file:

terminal
Loading...

IMPORTANT: Wait for Completion

Deep research takes 3-10 minutes to complete. The script will:

  1. Stream thinking/progress updates in real-time
  2. Automatically retry and poll for results if needed
  3. Save the final report to a file when complete

DO NOT interrupt or check on the research while it's running. The script handles all retries and status checking internally.

Examples

terminal
Loading...

Options

OptionShortDefaultDescription
--output-oauto-generatedOutput file path
--timeout-t600Maximum time in seconds (default: 10 min)
--verbose-vfalseEnable verbose debug logging
--systemnoneCustom system instruction

Research Process

The script uses Google's Deep Research agent which:

  1. Starts an interaction: Creates a research session with the topic
  2. Streams thinking summaries: Real-time progress as the agent researches
  3. Delivers results: Final report with sources and citations
  4. Auto-recovers: Handles network issues and retries automatically

Output

The script generates a markdown report containing:

  • Comprehensive research findings
  • Source URLs for verification (in Sources section)

By default, reports are saved to /workspace/deep_research/ with auto-generated filenames.

IMPORTANT: Enhance Reports with Images

After the research completes, you MUST enhance the report with generated images:

  1. Generate a title image for the report header
  2. Generate 1-2 concept visualizations for key findings
  3. Insert the images into the report using sed

Example workflow after research completes:

terminal
Loading...

Images should be saved to /workspace/generated_images/ and referenced with relative paths from the report.

Streaming Events

The script outputs JSON events for real-time UI updates:

  • {"event": "start", "topic": "..."} - Research started
  • {"event": "thinking", "summary": "..."} - Progress update
  • {"event": "content", "text": "..."} - Report content
  • {"event": "complete", "report_file": "...", "sources_count": N} - Done

Requirements

  • GEMINI_API_KEY environment variable must be set
  • Python 3.10+ with requests package installed

Notes

  • Research typically takes 3-10 minutes depending on topic complexity
  • The script blocks until completion - do not interrupt
  • Results include sources for verification
  • If research fails, the script will clearly indicate this