菜单
Craft and launch a Gemini Deep Research session via a custom Gem. Use when: @dp, @deep-research, research sprint, investigate topic, market research, technical deep-dive.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | gemini-deep-research-browser |
| alias | @dp |
| description | Craft and launch a Gemini Deep Research session via a custom Gem. Use when: @dp, @deep-research, research sprint, investigate topic, market research, technical deep-dive. |
| risk | safe |
| source | K3 Agentic Skills |
| date_added | 2026-05-12 |
Alias: @dp [topic]
Crafts and launches a Gemini Deep Research session using a custom Deep Research Architect Gem that you own.
https://gemini.google.com/gems/createDeep Research ArchitectYou are a Deep Research Query Architect. Your job is to transform vague research topics into high-precision Gemini Deep Research queries.
When a user gives you a topic:
1. Identify the core question and 2-3 critical sub-questions
2. Add year scope (current year), domain constraints, and format requests (tables, code examples, comparisons)
3. Recommend source type:
- Technical/general → No sources needed (web-only)
- User-provided docs → Google Drive attachment
- Prior research notes → NotebookLM attachment
4. Output:
a. The paste-ready research query (3-5 sentences, specific)
b. Source recommendation with reasoning
c. A 3-item plan checklist of what good research will cover
Keep queries tight. Vague queries produce vague reports.
@dp endpoint.Expand the user's topic:
Open your Gem URL and drop the topic in. It returns a paste-ready research query + plan checklist.
https://gemini.google.com/app+ button left of input → Tools → Deep researchWhen research completes:
.docxpython -m markitdown report.docx > report.md| Action | Where |
|---|---|
| Activate Deep Research | + button → Tools → "Deep research" |
| Attach Google Drive | Sources button → Drive |
| Edit research plan | "Edit plan" before "Start research" |
| Export report | Share icon → Export to Docs |
@dp Gemini 3.1 model cascading: Flash-Lite drafter → Flash linter → Pro critic — ADK LoopAgent cost and quality tradeoffs 2026
@dp thinking_level None/Low/Medium/High in Gemini 3.1 — token cost, latency, quality per level for multi-agent pipelines
@dp Gemini Batch Mode 50% discount — async batch processing in ADK for non-realtime generation
@dp ARC-AGI-2 benchmark results — what long-horizon agentic tasks Gemini 3.1 Pro handles vs fails at
@dp sqlite-vec vs pgvector for local vector memory at 10k documents — embedding strategies 2026
@dp Gemini Embedding 2 task prefix patterns vs legacy task_type — migration guide and pitfalls
@dp MRL truncation quality vs storage tradeoffs — 256/768/1536/3072 dims benchmarks 2026
@dp Jules GitHub agent autonomous PR workflows — deduplication, branch hygiene, CI/CD best practices 2026
@dp Docker dynamic MCP server pattern — runtime secret injection from GCP Secret Manager
@dp ADK DatabaseSessionService SQLite session persistence — call replay and failure logging patterns
@dp [your market] competitive landscape 2026 — pricing, features, differentiators comparison table
@dp [your tech stack] security audit patterns — OWASP Top 10 coverage and remediation 2026
Surfaces risks, blind spots, and opportunities the operator didn't ask about. Runs after plans are drafted but before execution. Use when reviewing implementation plans, launching features, or making architectural decisions.
Google Agent Development Kit (ADK) for Python — verified API patterns for building production-grade multi-agent systems. Covers Workflow graphs (ADK 2.0 Beta), LlmAgent, SequentialAgent, LoopAgent, CustomAgent (BaseAgent), MCP integration via McpToolset, FunctionTool, ToolContext state, session management, context compaction, callbacks (before/after agent/model/tool), human input via RequestInput, and the App wrapper. Use when: ADK agent, workflow graph, multi-agent, MCP toolset, HITL, context compaction, session state, agent pipeline.
Gemini Embedding 2 — Google's natively multimodal embedding model (GA April 2026). Maps text, images, video, audio, and PDFs into a unified 3072-dim vector space with MRL support (truncate to 128-3072). Use when: embeddings, semantic search, RAG, clustering, classification, multimodal retrieval, vector database, MRL truncation, cosine similarity, narrative memory, code retrieval, cross-modal search, agentic RAG.
Autonomous HestiaCP server management with AI agents. Two-brain Orchestrator+Sentinel architecture for VPS monitoring, troubleshooting, security hardening, PHP-FPM tuning, Exim mail forensics, and post-compromise detection. Use when: HestiaCP, VPS, server monitoring, DevOps agent, Linux admin, Nginx/Apache, PHP-FPM, Exim, Fail2Ban, ClamAV, SSL, backup.
You are a cognitive architect who understands that memory makes agents intelligent. You've built memory systems for agents handling millions of interactions. You know that the hard part isn't storing - it's retrieving the right memory at the right time.
Build secure, scalable authentication and authorization systems using industry-standard patterns and modern best practices.