zoharbabin/due-diligence-agents repository overview

⁠Due Diligence Agents

Legal flags a risk. Finance flags another. We connect and cite. Open-source forensic M&A due diligence — 13 AI agents read your entire data room across 9 specialist domains (Legal, Finance, Commercial, ProductTech, Cybersecurity, HR, Tax, Regulatory, ESG), cross-reference the findings no single reviewer connects, and trace every one to an exact page and verbatim quote.

📄 Documentation⁠ · 📊 Sample Report⁠ · 💻 Source Code⁠

⁠Quick Start

docker pull zoharbabin/due-diligence-agents:latest

# Generate a deal configuration
docker run --rm \
  -e ANTHROPIC_API_KEY="$ANTHROPIC_API_KEY" \
  -v ./data_room:/data \
  zoharbabin/due-diligence-agents auto-config "Buyer Corp" "Target Inc" --data-room /data

# Run the full analysis pipeline
docker run --rm \
  -e ANTHROPIC_API_KEY="$ANTHROPIC_API_KEY" \
  -v ./data_room:/data \
  -v ./deal-config.json:/workspace/deal-config.json \
  zoharbabin/due-diligence-agents run /workspace/deal-config.json

Output appears in your data room at _dd/forensic-dd/runs/latest/report/.

⁠What It Does

⁠What Gets Analyzed

• Legal — Change of control, anti-assignment, termination, IP ownership, indemnification, liability caps
• Finance — Revenue cross-referencing, unit economics (CAC/LTV/NRR), cost structure, projections
• Commercial — Renewal mechanics, customer concentration, SLAs, pricing models, MFN clauses
• ProductTech — DPA analysis, security certifications, technical SLAs, migration complexity
• Cybersecurity — Security governance, incident history, vulnerability management, disaster recovery
• HR — Compensation, key talent retention, labor compliance, workforce classification
• Tax — Transfer pricing, NOL/tax attributes, deal structure, income tax compliance
• Regulatory — License transferability, antitrust, data privacy, AML/sanctions
• ESG — Environmental contamination, climate risk, supply chain sustainability

⁠Output

• Interactive HTML report — Go/No-Go verdict, executive narrative, severity filtering, cross-domain synthesis
• 16-sheet Excel report — structured findings for downstream modeling
• Per-subject JSON — every finding with severity, citations, and cross-references

⁠Architecture

Data Room → Python Orchestrator (38 steps, 5 blocking gates)
         → 9 Specialist Agents (parallel) + Judge + Cross-Domain Analysis
         → Merge & Audit (dedup, citation verification, 31 QA checks)
         → Executive Synthesis (Go/No-Go signal)
         → HTML + Excel + JSON output

⁠Environment Variables

*Either Anthropic API key or AWS Bedrock credentials required.

⁠Docker Compose Example

services:
  dd-agents:
    image: zoharbabin/due-diligence-agents:latest
    environment:
      - ANTHROPIC_API_KEY=${ANTHROPIC_API_KEY}
    volumes:
      - ./data_room:/data
      - ./deal-config.json:/workspace/deal-config.json
    command: run /workspace/deal-config.json

⁠Tags

• latest — most recent stable release
• x.y.z — pin a specific release (matches the PyPI version; see Releases⁠)
• x.y — latest patch within a minor series

⁠Security & Privacy

• Local execution — documents only leave your machine as API calls to your LLM provider
• No telemetry — no phone-home, no usage data collection
• Read-only — never modifies files in your data room
• No persistent credentials — API keys read from environment, never stored in output

⁠Links

• Documentation: https://zoharbabin.com/due-diligence-agents/⁠
• Sample Report: https://zoharbabin.com/due-diligence-agents/sample-report/⁠
• Source Code: https://github.com/zoharbabin/due-diligence-agents⁠
• PyPI: https://pypi.org/project/dd-agents/⁠
• Issues: https://github.com/zoharbabin/due-diligence-agents/issues⁠

⁠License

Apache 2.0 — free for commercial use.