Financial Deep Research

Financial Deep Research

<!-- STATIC CONTEXT BLOCK START - Optimized for prompt caching --> <!-- All static instructions, methodology, and templates below this line --> <!-- Dynamic content (user queries, results) added after this block -->

Core System Instructions

Purpose: Deliver citation-backed, verified financial research reports through 8-phase pipeline (Scope > Plan > Retrieve > Triangulate > Synthesize > Critique > Refine > Package) with financial source credibility scoring, regulatory compliance tracking, and progressive context management.

Financial Focus: This skill specializes in:

• Market analysis and investment research
• Due diligence and competitive benchmarking
• Regulatory compliance and risk assessment
• Financial modeling support and valuation analysis
• Earnings analysis and financial statement review
• Sector/industry deep dives

Context Strategy: This skill uses 2025 context engineering best practices:

• Static instructions cached (this section)
• Progressive disclosure (load references only when needed)
• Avoid "loss in the middle" (critical info at start/end, not buried)
• Explicit section markers for context navigation

Decision Tree (Execute First)

Request Analysis
|-- Simple stock quote? -> STOP: Use WebSearch, not this skill
|-- Basic company lookup? -> STOP: Use WebSearch, not this skill
|-- Debugging code? -> STOP: Use standard tools, not this skill
+-- Complex financial analysis needed? -> CONTINUE

Mode Selection
|-- Quick market check? -> quick (3 phases, 2-5 min)
|-- Standard analysis? -> standard (6 phases, 5-10 min) [DEFAULT]
|-- Investment decision? -> deep (8 phases, 10-20 min)
|-- Due diligence/M&A? -> ultradeep (8+ phases, 20-45 min)

Execution Loop (per phase)
|-- Load phase instructions from [methodology](./reference/methodology.md#phase-N)
|-- Execute phase tasks
|-- Spawn parallel agents if applicable
+-- Update progress

Validation Gate
|-- Run `python scripts/validate_report.py --report [path]`
|-- Pass? -> Deliver
+-- Fail? -> Fix (max 2 attempts) -> Still fails? -> Escalate

Workflow (Clarify > Plan > Act > Verify > Report)

AUTONOMY PRINCIPLE: This skill operates independently. Infer assumptions from query context. Only stop for critical errors or incomprehensible queries.

1. Clarify (Rarely Needed - Prefer Autonomy)

DEFAULT: Proceed autonomously. Derive assumptions from query signals.

ONLY ask if CRITICALLY ambiguous:

• Query is incomprehensible (e.g., "analyze the thing")
• Contradictory requirements (e.g., "quick 50-source ultradeep analysis")
• Critical compliance/regulatory scope unclear

When in doubt: PROCEED with standard mode. User will redirect if incorrect.

Default assumptions:

• Company analysis -> Assume investor/analyst audience
• Sector query -> Assume comprehensive market view needed
• Valuation query -> Assume institutional-quality analysis
• Regulatory query -> Assume US jurisdiction unless specified
• Standard mode is default for most queries

2. Plan

Mode selection criteria:

• Quick (2-5 min): Market snapshot, earnings preview, quick check
• Standard (5-10 min): Most analysis, balanced depth/speed [DEFAULT]
• Deep (10-20 min): Investment decisions, detailed due diligence
• UltraDeep (20-45 min): M&A due diligence, comprehensive sector analysis

Announce plan and execute:

• Briefly state: selected mode, estimated time, number of sources
• Example: "Starting standard mode financial research (5-10 min, 15-30 sources)"
• Proceed without waiting for approval

3. Act (Phase Execution)

All modes execute:

• Phase 1: SCOPE - Define financial analysis boundaries (method)
• Phase 3: RETRIEVE - Parallel financial data gathering (5-10 concurrent searches + agents) (method)
• Phase 8: PACKAGE - Generate report using template

Standard/Deep/UltraDeep execute:

• Phase 2: PLAN - Financial research strategy formulation
• Phase 4: TRIANGULATE - Verify 3+ sources per financial claim
• Phase 4.5: OUTLINE REFINEMENT - Adapt structure based on evidence (WebWeaver 2025) (method)
• Phase 5: SYNTHESIZE - Generate investment insights

Deep/UltraDeep execute:

• Phase 6: CRITIQUE - Risk analysis and bear case
• Phase 7: REFINE - Address gaps, strengthen thesis

Critical: Avoid "Loss in the Middle"

• Place key findings at START and END of sections, not buried
• Use explicit headers and markers
• Structure: Summary > Details > Conclusion (not Details sandwiched)

Progressive Context Loading:

• Load methodology sections on-demand
• Load template only for Phase 8
• Do not inline everything - reference external files

Anti-Hallucination Protocol (CRITICAL for Financial Data):

• Source grounding: Every financial claim MUST cite a specific source immediately [N]
• Clear boundaries: Distinguish between FACTS (from filings/data) and ANALYSIS (your interpretation)
• Explicit markers: Use "According to [1]..." or "[1] reports..." for source-grounded statements
• No speculation without labeling: Mark inferences as "This suggests..." not "Data shows..."
• Verify before citing: If unsure whether source actually says X, do NOT fabricate citation
• When uncertain: Say "No sources found for X" rather than inventing references
• Financial precision: Always include specific numbers, dates, and currency when available

Parallel Execution Requirements (CRITICAL for Speed):

Phase 3 RETRIEVE - Mandatory Parallel Financial Search:

1. Decompose query into 5-10 independent search angles before ANY searches
2. Launch ALL searches in single message with multiple tool calls (NOT sequential)
3. Quality threshold monitoring for FFS pattern:

• Track source count and avg credibility score
• Proceed when threshold reached (mode-specific, see methodology)
• Continue background searches for additional depth

1. Spawn 3-5 parallel agents using Task tool for deep-dive investigations

Financial Search Decomposition Strategy:

[Single message with 8+ parallel tool calls]
WebSearch #1: Company fundamentals + recent filings
WebSearch #2: Earnings/financial performance
WebSearch #3: Industry/sector analysis
WebSearch #4: Competitive landscape
WebSearch #5: Regulatory/compliance news
WebSearch #6: Analyst ratings/price targets
WebSearch #7: Risk factors/bear case
WebSearch #8: Recent news + catalysts
Task agent #1: SEC filing deep dive (10-K, 10-Q analysis)
Task agent #2: Financial statement analysis
Task agent #3: Industry comparison/benchmarking

4. Verify (Always Execute)

Step 1: Citation Verification (Catches Fabricated Sources)

python scripts/verify_citations.py --report [path]

Financial-Specific Checks:

• SEC filing references (verify EDGAR links)
• Financial data accuracy (cross-check key metrics)
• Date accuracy (earnings dates, filing dates)
• Flags suspicious entries (future financials, impossible metrics)

If suspicious citations found:

• Review flagged entries manually
• Remove or replace fabricated sources
• Re-run until clean

Step 2: Structure & Quality Validation

python scripts/validate_report.py --report [path]

9 automated checks (financial-enhanced):

1. Executive summary length (50-250 words)
2. Required sections present (+ recommended: Risk Factors, Valuation)
3. Citations formatted [1], [2], [3]
4. Bibliography matches citations
5. No placeholder text (TBD, TODO)
6. Word count reasonable (500-10000)
7. Minimum 10 sources
8. No broken internal links
9. Financial data consistency (dates, currencies, units)

If fails:

• Attempt 1: Auto-fix formatting/links