Menu
Conduct thorough research using multi-angle queries, reading multiple sources, and cross-referencing.
Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.
Baseado na classificação ocupacional SOC
Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction.
Translate a web page while preserving its original DOM structure, CSS, and images. Generates a self-contained bilingual preview HTML with toggle controls. Use this skill whenever the user shares a URL and asks for translation, translated preview, bilingual view, or when the channel topic instructs to use translate-preview. Do NOT write your own translation scripts — always use this skill's reference files.
Job scheduling. Manage recurring tasks using MCP tools.
Generate a prioritized daily briefing with calendar, email, tasks, and yesterday's activity. Designed for cron execution. Trigger with "morning briefing", "daily brief", "start my day".
Self-management via geminiclaw_admin MCP tool. Config, skills, upgrade, sessions.
Proactively explore the web for topics matching the user's interests and share findings conversationally. Designed for cron execution with Flash model.
| name | deep-research |
| description | Conduct thorough research using multi-angle queries, reading multiple sources, and cross-referencing. |
| enabled | true |
A skill for systematically investigating topics that cannot be answered with a single search or that require cross-referencing multiple sources.
Use this skill when the following apply:
If the scope is ambiguous, ask questions before proceeding:
Declare the angles to cover to the user before execution:
Research axes:
- Official information & specs
- User reviews & opinions
- Comparison with competitors & alternatives
- Latest developments (date-stamped)
- Community & expert discussions
Execute independent queries for each axis. Do not repeat the same angle.
| Axis | Query Example |
|---|---|
| Official / authoritative sources | {topic} official documentation specs |
| Reviews & opinions | {topic} review opinion site:reddit.com OR news.ycombinator.com |
| Comparison & benchmarks | {topic} vs {competitor} comparison 2026 |
| Latest information | {topic} 2026 latest update |
| Community | {topic} issues drawbacks pitfalls |
Prohibited: Drawing conclusions from a single search. Answering based on snippets alone without reading actual pages.
Search result snippets are only summaries. Read the full text of the top 3-5 results using web_fetch or agent-browser.
agent-browserweb_fetchCross-check important claims (prices, specs, ratings, figures) with 2 or more sources.
Save to memory/research-YYYY-MM-DD-{topic}.md (get the date from geminiclaw_status):
# Research: {topic}
Date: YYYY-MM-DD
## Summary
(1-3 sentences)
## Findings
### {Axis 1}
...
### {Axis 2}
...
## Conclusion
(Judgment / recommendation / caveats)
## Sources
- [Title](URL) — Retrieved: YYYY-MM-DD
Report to the user in the following format:
## Research Results: {topic}
### Summary
(2-3 sentence overview)
### Key Findings
- **{Axis}**: ...
- **{Axis}**: ...
### Conclusion
(Recommendation / judgment / caveats)
### Sources
- [Title](URL)