For Analysts and Policy / Risk Teams

One-minute route into Agenda Intelligence — an open-source AI-skill portfolio for structured strategic-risk analysis. For fund analysts, compliance and sanctions desks, journalists, advisory shops, and AI builders embedding regional intelligence into agents.

Email for a call LinkedIn Main profile

What Agenda Intelligence is

A four-repo system for AI agents producing auditable strategic-risk analysis — not generic geopolitical summaries.

• Mechanism-first reasoning. Primary driver → transmission channel → exposure map → role-based actions → trigger points. Not “monitor closely”.
• Per-claim provenance (Axis A / Axis B). Every factual claim carries a source-type tag ([primary] / [secondary] / [user-provided] / [inference] / [analyst-judgment]) and optional action flags ([verify], [stale-risk: YYYY-MM]).
• Four labelled evidence modes. live-source-backed, user-provided sources, illustrative source packet, reasoning-only. Each example states its mode openly.
• Honesty rules in code. CI validators block unsupported claims and require explicit disclosure when no production-use record exists.

The four repos

Role	Repo	What it covers
Horizontal reasoning skill	Global Think Tank Analyst	Cross-domain policy-risk memo structure: sanctions, regulatory, geopolitical, trade, scenario and red-team modes.
Vertical: Central Asia + Caspian	central-asia-caspian-hybrid-intelligence-skill	Sanctions, AML, banking, beneficial-ownership opacity, corridor logistics and energy risk across Kazakhstan, Uzbekistan, Azerbaijan, Turkmenistan, Kyrgyzstan, Tajikistan and the Caspian.
Vertical: Gulf + Middle East	gulf-middle-east-hybrid-intelligence-skill	Iran sanctions and OFAC SDN exposure, GCC banking and sovereign wealth (PIF, ADIA, Mubadala, QIA, KIA), energy and OPEC+ dynamics, Hormuz / Bab-el-Mandeb / Red Sea maritime chokepoint risk.
Evidence + audit layer	Agenda Intelligence MD	Validation, JSON schemas, evidence audit, scoring rubrics, CLI and MCP tooling so strategic-intelligence agent output can be audited rather than only read.

Who this is for

• Fund and corporate analysts in energy, logistics, banking and trade with Central Asia, Caspian, Iran, GCC or maritime exposure.
• Sanctions, AML and compliance desks that need structured reasoning before screening, not in place of it.
• Journalists and researchers covering sanctions enforcement, sovereign wealth, energy markets, corridor and dual-use trade.
• Policy and government advisors working on regional risk briefs.
• AI builders and ops teams embedding strategic-risk reasoning into agents, copilots or internal tools.

What this is not

• Not a sanctions or AML screening tool.
• Not legal, compliance, investment, or intelligence advice.
• Not a live source retriever or factuality verifier.
• Not a production-grade risk-control system.
• Not a benchmark. No accuracy scores are claimed.

Case studies

• Global Think Tank Analyst — horizontal reasoning skill
• Central Asia + Caspian Hybrid Intelligence Skill — vertical specialist
• Gulf + Middle East Hybrid Intelligence Skill — vertical specialist
• Agenda Intelligence MD — evidence and audit layer
• Eval run: Agenda Intelligence MD, May 2026 — N=3 self-scored run with full caveats. Not a benchmark.

How to engage

• Read first. The repos and case studies are public. Each repo has a Try this prompt block you can paste into Claude or Codex against your own scenario.
• Review collaboration. All three skill repos publish a STATUS.md with a “Bar 2” external-review section. If you are a sanctions, AML, energy-trading, shipping, Gulf-banking, Central-Asia-corridor or Iran-watcher practitioner and would consider reviewing an example, open an issue on the relevant repo or email below.
• Bespoke analysis. For non-public scenarios — desk briefings, region-specific risk memos, advisory under retainer — email with your decision context, geography and time horizon.

Contact

• Email: vassiliy.lakhonin@gmail.com
• LinkedIn: linkedin.com/in/vassiliy-lakhonin
• GitHub: github.com/vassiliylakhonin
• Location: Almaty, Kazakhstan (UTC+5)
• Response time: usually within 24 hours.