Cross-Border Oil Contract & Sanctions Multi-Agent Graph with LangChain and Across AI
Executing crude export mechanisms involving both local and foreign trading entities requires multi-layered legal and international compliance verification. Using LangChain and LangGraph alongside Across AI multi-agent communication networks, build a stateful execution graph that cross-references deal terms against international sanctions, trade embargos, and local banking clearing conditions. The system coordinates three specialized sub-agents: Legal Document Parser, Sanctions Screening Agent, and Financial Settlement Verifier. The graph enforces checkpointing, conditional state transitions, and automatic escalation.
What you are building
The core problem, expected build, and operating context for this challenge.
Construct a multi-agent LangChain and LangGraph workflow integrated with Across AI for automated crude export compliance audits.
How work is evaluated
Evaluates contract audit graph accuracy against sanctioned counterparty profiles.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
trade_clearance_check
Ensure clean entity is approved without false flags.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
audit_precision
Precision score on flagged compliance violations. • target: 0.99 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Build stateful multi-agent graphs using LangGraph and LangChain.
Integrate Across AI agent-to-agent communication protocols.
Implement automated compliance escalation and risk-scoring logic.
Reference links and supporting material
Synthetic legal crude contracts and associated OFAC/UN sanctions watchlist entries.
Challenge setup
Evaluates contract audit graph accuracy against sanctioned counterparty profiles.
Challenge input
JSON containing deal_id, counterparty_name, bank_swift, origin_country
LangChain
Industry standard framework for chaining LLM operations.
Across AI
Enables standardized inter-agent communication channels.
Evaluated output
JSON with clear_trade, risk_score, audit_reason
- Ensure clean entity is approved without false flags.
- Precision score on flagged compliance violations. • target: 0.99 • range: 0-1
- Benchmark: LangGraph-Compliance-Bench
- Audit Precision target: 0.99
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- LangChain · Required
- Langchain · Optional
- Aim · Optional
- LangChain · Required
- Langchain · Optional
Evaluation contract
- trade_clearance_check · Weight 1
- audit_precision · Weight 1
Metrics
- audit_precision · Target 0.99
Benchmarks
- LangGraph-Compliance-Bench
- security
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON containing deal_id, counterparty_name, bank_swift, origin_country
Output: JSON with clear_trade, risk_score, audit_reason
Execution harness: Python
Evaluation tools
- LangChain · Framework Core
- Across AI · Agent Communication
Recipe freeze marker
No recipe freeze time is recorded.