Skip to content

Deep Research

Runs multi-source research with parallel sub-agents and writes a cited report.

Stars
13,743
Licence
MIT
Files
4 files

What it does

When you ask for a thorough, multi-source investigation, the agent plans several independent research angles, has parallel sub-agents search and collect evidence for each, checks for gaps, and then writes one cited report saved to a research folder. It is meant for questions that need many sources, not quick lookups.

Great for

  • People who need a sourced briefing on a topic before making a decision
  • Analysts comparing many options across several fields
  • Users who want a report with citations and noted open questions

Skip it if

  • Simple factual questions, since the skill itself says a single search is enough and this adds cost and time
  • Academic literature surveys, which the skill directs to a different auto-research skill
  • Situations needing verified, primary-source accuracy without review, since the report relies on web search results that can be wrong or stale

Try asking

  • “Deep research on the current state of solid-state battery technology with a cited report”
  • “Compare 20 open-source note-taking apps across pricing, platforms and privacy”
  • “帮我全面研究一下 remote work productivity studies from the last two years”
  • “Exhaustive research on counter-arguments to a proposed city congestion tax”
Step-by-step guide: how to use Deep ResearchInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 4 original files, the MIT licence and an attribution note.

    deep-research.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip deep-research.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip deep-research.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • State the depth you want (quick, standard or deep) and the audience or decision at stake so the agent does not need to ask many questions
  • For deep mode or contested topics, review the proposed research angles when the agent shows them before budget is spent
  • Check the cited sources yourself for any claim that matters, especially those marked [single source] or [speculative]

What's inside

  • reference/report.md
  • reference/sources.md
  • reference/subagent-prompt.md
  • SKILL.md

Redistributed unmodified from XiaomiMiMo/MiMo-Code under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.