How sherlinzenith/agentic-dd works
This is an agentic due-diligence platform that turns workspace documents into traceable findings for human review and reporting, with a separate evidence-grounded Q&A path. The documented workflow spans extraction, classification, planning, specialist analysis, verification, review, and reporting. The graph reflects the README and the supplied excerpts/index; several workflow internals and UI/API wiring are only partially shown, so uncertain links are omitted.
Main entry points: orchestrator.py DD orchestrator · document_processor.py Document extraction · finding_store.py Findings and review store · main.py FastAPI endpoints
Components 21
Application entry
- Streamlit UI
streamlit_app.py - FastAPI endpoints
app/main.py - UI engine service
ui/service.py
Document intelligence
- Document extraction
processing/document_processor.py - Workstream classification
processing/document_classifier.py - Document inventory
agents/inventory.py - Readiness checks
agents/readiness.py
Investigation workflow
- Investigation planning
agents/lead_agent.py - Job lifecycle
workflow/dd_graph.py - DD orchestrator
workflow/orchestrator.py - Financial agent
agents/financial_agent.py - Legal agent
agents/legal_agent.py - Cross-domain analysis
agents/cross_domain.py - Evidence verification
agents/evidence_verify.py
Findings and review
- Findings and review store
findings/finding_store.py - DD report generation
reports/report_generator.py
Workspace Q&A
- Evidence-grounded Q&A
qa/qa_service.py - Workspace retrieval
qa/retriever.py - Q&A data models
qa/models.py
Other components
- Reviewer
- VDR workspace
Connections 29
- VDR workspace → Document extraction: supplies documents
README.md - Document extraction → Workstream classification: provides text
app/main.py - FastAPI endpoints → Document extraction: loads documents
app/main.py - FastAPI endpoints → Workstream classification: classifies documents
app/main.py - Streamlit UI → UI engine service: uses service
streamlit_app.py - UI engine service → Document extraction: loads documents
ui/service.py - UI engine service → Document inventory: organizes documents
ui/service.py - UI engine service → Readiness checks: checks readiness
ui/service.py - FastAPI endpoints → Job lifecycle: starts and resumes jobs
app/main.py - Job lifecycle → Document extraction: loads documents
workflow/dd_graph.py - Job lifecycle → Investigation planning: validates plan types
workflow/dd_graph.py - Job lifecycle → DD orchestrator: builds workflow
workflow/dd_graph.py - Investigation planning → DD orchestrator: plans workstreams
README.md - DD orchestrator → Financial agent: dispatches financial work
workflow/orchestrator.py - DD orchestrator → Legal agent: dispatches legal work
workflow/orchestrator.py - DD orchestrator → Cross-domain analysis: runs cross-domain analysis
workflow/orchestrator.py - DD orchestrator → Evidence verification: verifies findings
README.md - Financial agent → Findings and review store: saves findings
agents/financial_agent.py - Evidence verification → Findings and review store: supports findings
README.md - Reviewer → Findings and review store: approves or rejects
README.md - FastAPI endpoints → Findings and review store: records review
app/main.py - Job lifecycle → Findings and review store: resets job findings
workflow/dd_graph.py - FastAPI endpoints → DD report generation: generates report
app/main.py - Findings and review store → DD report generation: provides findings
README.md - Streamlit UI → Evidence-grounded Q&A: asks workspace questions
streamlit_app.py - Evidence-grounded Q&A → Workspace retrieval: retrieves evidence
qa/qa_service.py - Evidence-grounded Q&A → Q&A data models: returns answer models
qa/qa_service.py - Workspace retrieval → Q&A data models: returns retrieved chunks
qa/retriever.py - Document extraction → Evidence-grounded Q&A: supplies workspace documents