Gemini Deep Research Brief skill for Claude and Codex
Scopes an AI deep research run with a sharp question, ranked sub-questions and source rules, so the report supports a decision.
Install it once and Claude or Codex runs it for you. When your request matches, your AI loads the skill, asks for anything it is missing, and works through the template itself. Prefer to fill it in by hand? See the full prompt template.
Download the skill (.zip) · View SKILL.md
When it is used
Use when someone is about to start an AI deep research run and needs a brief that fixes the question, scope, sources and report structure. Do not use for running the research itself or summarising a single document. Related topics: AI, Research, Productivity.
What it will ask you for
- Research question in one line (free text, e.g. "Which vector database suits a 10 million document archive")
- Why this is needed (longer free text)
- Type of research (choose one: Market landscape, Technical comparison, Buying decision, Trend analysis, Academic overview)
- Must-cover subtopics or competitors (longer free text)
- Time range that matters (free text, e.g. "last 12 months")
- Source and reporting rules (choose any: Prefer primary sources, Include pricing data, Include expert opinions, Flag conflicting claims, Note confidence per claim)
- Structure (choose one: Executive summary first, Comparison table first, Chronological)
- Maximum length (free text, e.g. "2000 words")
How it works
- Rewrite the research question so it names the decision it feeds and the time range that matters.
- Break it into sub-questions ordered by their effect on the decision, each answerable with evidence.
- Set source guidance from the chosen rules, including how conflicting claims and confidence are reported.
- Define scope exclusions so the run does not spread into adjacent topics.
- Lay out the report with exact section headings matching the chosen structure and length limit.
What you get
A brief with the refined question, ranked sub-questions, source guidance, scope exclusions, and a headed report outline within the length limit.
What it watches out for
- Keeping the question broad, which returns an encyclopedic overview.
- Omitting the decision context, so the report cannot weigh what matters.
- Asking for confidence notes without saying how to express them.
How to install
Claude.ai and the Claude desktop app
- Download the skill .zip.
- Open Settings, then Capabilities, and find Skills.
- Choose Upload skill and select the .zip.
- Ask Claude for help with the task. It loads the skill when your request matches.
Claude Code
- Download the skill .zip.
- Unzip it into ~/.claude/skills/ (or .claude/skills/ inside a project).
- Start a new session. Claude Code loads the skill when your request matches.
Codex
- Download the skill .zip.
- Unzip it into ~/.agents/skills/ (or .agents/skills/ inside a project).
- Codex picks it up on its own. Type $ and the skill name to use it, or just ask for the task.
Cursor
- Download the Cursor rule (.mdc).
- Save it in .cursor/rules/ inside your project.
- Cursor's agent applies the rule when your request matches its description.
AGENTS.md (Codex, Cursor, and other coding agents)
- Copy the AGENTS.md section.
- Paste it into AGENTS.md at the root of your repository, creating the file if needed.
- Coding agents that read AGENTS.md follow it when your request fits.