Skip to main content

gemini-web-research

Use when the user asks to delegate substantial web research to Gemini CLI or agy. Do not invoke for ordinary research requests or a quick search.

ソース情報

リポジトリ
zazencodes/skills
ソースの最終更新活動
2026年10月5日 05:09
検出された SKILL.md の言語
英語
スター
0
フォーク
0

gemini-web-research: coordinate cited research in Gemini sessions

Delegate substantial web research to interactive Gemini sessions in tmux panes. The coordinating agent reviews cited reports, follows up in the sessions that hold the relevant context and synthesizes the findings.

Prerequisites

The source uses agy, tmux and existing model access. The user must explicitly ask to delegate research to Gemini or agy.

How to use

Provide the research question and scope. Split it into independent areas, launch a session for each, read the reports and send targeted follow-ups to the same pane. Before using consequential findings, have the worker open and verify the cited sources.

Limitations

The source excludes ordinary research requests and quick searches. A failed or closed session is reported rather than silently switching services. Reports and citations need review before they become the final synthesis.

インストール方法

デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。

ソースファイルを確認

インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。

ファイルエクスプローラー
4 ファイル

SKILL.md を表示中

SKILL.md
ソースの指示 · 読み取り専用プレビュー
name
gemini-web-research
description
Use when the user asks to delegate substantial web research to Gemini CLI or agy. Do not invoke for ordinary research requests or a quick search.
# Gemini Web Research Orchestrate web research through interactive `agy` sessions in tmux panes, then use their findings in the parent task. Invoke this skill only when the user asks to use Gemini or `agy` for research, including a clear request to delegate research to Gemini. Do not choose it merely because research is extensive. ## Roles You are the orchestrator; the `agy` sessions are workers. They do the expensive work: searching, opening pages, checking links, distilling. You decide what to ask, read their reports, spot gaps and contradictions, and send the next question. Do not run web searches or fetches yourself. If something needs checking, send it to a worker. Your context is for thinking and synthesis, not raw pages. ## Workflow 1. **Split the question.** Break the research into independent areas and launch one session per area, in parallel. Keep each first prompt light and narrow; depth comes from follow-ups. 2. **Read each report as it lands.** Look for what is missing, unverified, thin, out of scope, or in conflict with another report. 3. **Follow up in the same session.** For targeted digging on an area a session already covered (verify these links, go deeper on X, fill the Saturday gap), send the follow-up to that pane. The session keeps its context, so the follow-up can be short. 4. **Spin up new sessions** for unrelated areas, or when a session is gone or confused. 5. **Iterate** until the findings answer the question, then synthesize. ## Launch a session The user always works in tmux. Each session runs in a new pane so they can watch it: ```bash <this skill's directory>/scripts/agy-research.sh <scratch>/<area>-1.md 'Research <topic and scope>. Search the web and cite direct source URLs beside substantive claims. Distinguish publication dates from event dates. Flag uncertainty. Take no external actions and edit no files except the report. When finished, write the complete report in a single write to <scratch>/<area>-1.md.' ``` The launcher splits the current pane, starts `agy -i` with Gemini 3.8 Flash at high effort, prints `started in tmux pane %<id>` to stderr, waits until the report exists, then prints it. Run each launcher as its own background command, without redirecting stderr away, so every session notifies you separately when it finishes and its pane id stays in the output. Record which pane covers which area. If the model is unavailable, check `agy models`, choose the current Gemini Flash high model, and update both scripts. ## Follow up in a session ```bash <this skill's directory>/scripts/agy-followup.sh %<id> <scratch>/<area>-2.md 'Follow-up: <targeted question>. Write the complete answer in a single write to <scratch>/<area>-2.md.' ``` This types the prompt into the pane with `tmux send-keys`, waits for the new report, then prints it. The prompt must be a single line, and each follow-up needs a new report path. Run it in the background like a launch. Send follow-ups only to a session that has finished its last report. Panes stay open after the turn; the user closes them. If a pane closes before its report is written, the script exits with an error: report it rather than silently switching research services. ## Use the result Have a worker verify the key cited sources before you repeat consequential claims. Ask it to open each link and confirm it supports the claim. Correct or mark as unverified anything that fails. Return a synthesis with links, not raw Gemini reports.
GitHubで見る