feat: improve research agent prompt design and delegation strategy

- Add date formatting to RESEARCHER_INSTRUCTIONS placeholder
- Reorder workflow steps (plan before saving request)
- Add synthesis step for consolidating sub-agent findings
- Separate orchestrator and sub-agent instructions
- Remove RESEARCHER_INSTRUCTIONS from main agent (orchestrator only delegates)
- Simplify citation format and remove duplication
- Bias delegation strategy toward single sub-agent by default
- Only parallelize for explicit comparisons or clearly separated aspects
- Enhance verification step to check citations and structure

.gitignore

@@ -163,4 +163,7 @@ cython_debug/
 #  option (not recommended) you can uncomment the following to ignore the entire idea folder.
 #.idea/
 
-.langgraph_api/
+# Other 
+.langgraph_api/
+.vscode/
+.DS_Store

deep_research/agent.py

@@ -4,7 +4,10 @@ This module creates a deep research agent with custom tools and prompts
 for conducting web research with strategic thinking and context management.
 """
 
+from datetime import datetime
+
 from langchain.chat_models import init_chat_model
+from langchain_google_genai import ChatGoogleGenerativeAI
 from deepagents import create_deep_agent
 
 from research_agent.prompts import (
@@ -18,7 +21,10 @@ from research_agent.tools import tavily_search, think_tool
 max_concurrent_research_units = 3
 max_researcher_iterations = 3
 
-# Combine instructions
+# Get current date
+current_date = datetime.now().strftime("%Y-%m-%d")
+
+# Combine orchestrator instructions (RESEARCHER_INSTRUCTIONS only for sub-agents)
 INSTRUCTIONS = (
     RESEARCH_WORKFLOW_INSTRUCTIONS
     + "\n\n"
@@ -28,25 +34,21 @@ INSTRUCTIONS = (
         max_concurrent_research_units=max_concurrent_research_units,
         max_researcher_iterations=max_researcher_iterations,
     )
-    + "\n\n"
-    + "=" * 80
-    + "\n\n"
-    + RESEARCHER_INSTRUCTIONS
 )
 
 # Create research sub-agent
 research_sub_agent = {
     "name": "research-agent",
     "description": "Delegate research to the sub-agent researcher. Only give this researcher one topic at a time.",
-    "system_prompt": RESEARCHER_INSTRUCTIONS,
+    "system_prompt": RESEARCHER_INSTRUCTIONS.format(date=current_date),
     "tools": [tavily_search, think_tool],
 }
 
-# Initialize model
-model = init_chat_model(
-    model="anthropic:claude-sonnet-4-5-20250929",
-    temperature=0.0
-)
+# Model Claude 4.5
+# model = init_chat_model(model="anthropic:claude-sonnet-4-5-20250929", temperature=0.0)
+
+# Model Gemini 3 
+model = ChatGoogleGenerativeAI(model="gemini-3-pro-preview", temperature=0.0)
 
 # Create the agent
 agent = create_deep_agent(

deep_research/research_agent/prompts.py

@@ -4,11 +4,12 @@ RESEARCH_WORKFLOW_INSTRUCTIONS = """# Research Workflow
 
 Follow this workflow for all research requests:
 
-1. **Save the request**: Use write_file() to save the user's research question to `/research_request.md`
-2. **Plan**: Create a todo list with write_todos to break down the research into focused tasks
-3. **Research**: Delegate research tasks to sub-agents using the task() tool
-4. **Write Report**: Write a comprehensive final report to `/final_report.md` (see Report Writing Guidelines below)
-5. **Verify**: Read `/research_request.md` to confirm you've addressed all aspects of the original question
+1. **Plan**: Create a todo list with write_todos to break down the research into focused tasks
+2. **Save the request**: Use write_file() to save the user's research question to `/research_request.md`
+3. **Research**: Delegate research tasks to sub-agents using the task() tool - ALWAYS use sub-agents for research, never conduct research yourself
+4. **Synthesize**: Review all sub-agent findings and consolidate citations (each unique URL gets one number across all findings)
+5. **Write Report**: Write a comprehensive final report to `/final_report.md` (see Report Writing Guidelines below)
+6. **Verify**: Read `/research_request.md` and confirm you've addressed all aspects with proper citations and structure
 
 ## Research Planning Guidelines
 - Batch similar research tasks into a single TODO to minimize overhead
@@ -48,10 +49,9 @@ Simply list items with details - no introduction needed:
 - Each section should be comprehensive and detailed
 - Use bullet points only when listing is more appropriate than prose
 
-<Citation Rules>
-Use numbered citations throughout your report:
+**Citation format:**
 - Cite sources inline using [1], [2], [3] format
-- Assign each unique URL a single citation number
+- Assign each unique URL a single citation number across ALL sub-agent findings
 - End report with ### Sources section listing each numbered source
 - Number sources sequentially without gaps (1,2,3,4...)
 - Format: [1] Source Title: URL (each on separate line for proper list rendering)
@@ -62,7 +62,6 @@ Use numbered citations throughout your report:
   ### Sources
   [1] AI Research Paper: https://example.com/paper
   [2] Industry Analysis: https://example.com/analysis
-</Citation Rules>
 """
 
 RESEARCHER_INSTRUCTIONS = """You are a research assistant conducting research on the user's input topic. For context, today's date is {date}.
@@ -111,20 +110,24 @@ After each search tool call, use think_tool to analyze the results:
 </Show Your Thinking>
 
 <Final Response Format>
-When providing your final answer, use numbered citations and include a Sources section:
+When providing your findings back to the orchestrator:
 
-1. **Cite sources inline**: Use [1], [2], [3] format when referencing information
-2. **Sources section**: End with a numbered list of all sources
+1. **Structure your response**: Organize findings with clear headings and detailed explanations
+2. **Cite sources inline**: Use [1], [2], [3] format when referencing information from your searches
+3. **Include Sources section**: End with ### Sources listing each numbered source with title and URL
 
-Example format:
+Example:
+```
+## Key Findings
 
 Context engineering is a critical technique for AI agents [1]. Studies show that proper context management can improve performance by 40% [2].
 
 ### Sources
 [1] Context Engineering Guide: https://example.com/context-guide
 [2] AI Performance Study: https://example.com/study
+```
 
-This allows readers to verify information and explore topics further.
+The orchestrator will consolidate citations from all sub-agents into the final report.
 </Final Response Format>
 """
 
@@ -138,20 +141,28 @@ Your role is to coordinate research by delegating tasks from your TODO list to s
 
 ## Delegation Strategy
 
-**Simple queries** → 1 sub-agent:
-- "List the top 10 coffee shops in San Francisco"
-- "What is quantum computing?"
-- "Summarize the history of the internet"
+**DEFAULT: Start with 1 sub-agent** for most queries:
+- "What is quantum computing?" → 1 sub-agent (general overview)
+- "List the top 10 coffee shops in San Francisco" → 1 sub-agent
+- "Summarize the history of the internet" → 1 sub-agent
+- "Research context engineering for AI agents" → 1 sub-agent (covers all aspects)
 
-**Comparisons** → 1 sub-agent per element:
+**ONLY parallelize when the query EXPLICITLY requires comparison or has clearly independent aspects:**
+
+**Explicit comparisons** → 1 sub-agent per element:
 - "Compare OpenAI vs Anthropic vs DeepMind AI safety approaches" → 3 parallel sub-agents
 - "Compare Python vs JavaScript for web development" → 2 parallel sub-agents
 
-**Multi-faceted research** → 1 sub-agent per aspect:
-- "Research renewable energy: costs, environmental impact, adoption rates" → 3 parallel sub-agents
-- "Analyze the 2024 election: candidates, issues, polling" → 3 parallel sub-agents
+**Clearly separated aspects** → 1 sub-agent per aspect (use sparingly):
+- "Research renewable energy adoption in Europe, Asia, and North America" → 3 parallel sub-agents (geographic separation)
+- Only use this pattern when aspects cannot be covered efficiently by a single comprehensive search
 
-## Parallel Execution
+## Key Principles
+- **Bias towards single sub-agent**: One comprehensive research task is more token-efficient than multiple narrow ones
+- **Avoid premature decomposition**: Don't break "research X" into "research X overview", "research X techniques", "research X applications" - just use 1 sub-agent for all of X
+- **Parallelize only for clear comparisons**: Use multiple sub-agents when comparing distinct entities or geographically separated data
+
+## Parallel Execution Limits
 - Use at most {max_concurrent_research_units} parallel sub-agents per iteration
 - Make multiple task() calls in a single response to enable parallel execution
 - Each sub-agent returns findings independently