ABOUT · WARNING OF WAR
Clinical risk intelligence, rebuilt every three hours.
Warning of War is an open-access geopolitical & supply-chain risk monitor. It re-reads global open-source business news on a fixed cadence and decomposes what it finds into calibrated, comparable risk scores — across seven world regions, nine maritime choke points, and four industry verticals — so that risk professionals get a clinical second opinion without the noise.
What this is
Most geopolitical coverage is written to be read; Warning of War is written to be used. Every three hours an automated pipeline ingests curated business, energy, sanctions, commodity, and shipping feeds, then scores the pressure it finds on a fixed 1–10 scale per sector axis and a 0–100 composite per region. The same cycle re-assesses nine global maritime choke points and four industry hubs — maritime logistics, energy markets, commodities & raw materials, and macroeconomic impact — each decomposed into five operational axes and five industry verticals.
Five weekly report series — a Monday cross-domain outlook plus four deep dives — distil the live data into shareable scorecards, carousels, and newsletter editions, published under /reports/.
Everything is open: the scores, the briefs, the JSON feeds behind them, and the full methodology. Data is released under CC BY 4.0 — cite it, build on it, hold it to account.
Who it serves
- Enterprise risk & supply-chain teams — a fast, consistent read on where cross-region pressure is building before it reaches procurement and logistics.
- Institutional investors & analysts — a calibrated, history-tracked composite that can be compared across regions and through time.
- Maritime & logistics operators — choke-point status, lane-level disruption events, and sanctions-compliance signals in one place.
It is explicitly not a war-news feed. Coverage is framed strictly around operational, commercial, and financial impact — no casualty reporting, no humanitarian framing, no alarm language. That constraint is contractual to the pipeline, not stylistic: it is enforced in the model's instructions on every call.
How it works
The pipeline is deliberately simple and fully auditable — the complete source code is public at github.com/vency964/warningofwar:
- Ingest — curated RSS pools per region and per industry hub, biased toward business desks, sanctions trackers, and shipping authorities; de-noised and de-duplicated before any model call.
- Decompose — Anthropic's Claude Sonnet 4.6 scores each region and hub against a strict, strongly-typed output schema with a fixed analyst role and categorical brand-safety rules.
- Score — fixed, published weights combine the sector axes into each composite; no hidden adjustments, no manual overrides.
- Publish — the entire site is prerendered as static HTML and deployed on every cycle; the JSON feeds the pages render from are public.
The full reference — axis definitions, scoring rubric, weights, limitations, and what each number does not mean — lives on the methodology page.
Who is behind it
Warning of War was founded in May 2026 by Ventsislav Georgiev and is independently owned and operated — no institutional backing, no sponsored placement, no paid rankings.
The scoring itself is AI-driven by design, and the site is transparent about that: every brief is generated by Claude Sonnet 4.6 against a fixed, human-maintained methodology. The founder maintains what the model cannot decide for itself — the source lists, the sector taxonomy, the scoring weights, the brand-safety constraints, and the editorial standards — and reviews the system's output and calibration on an ongoing basis.
Editorial standards & corrections
- Clinical voice. Operational, commercial, and financial framing only. Kinetic events are described by their effect on assets, lanes, prices, and operations.
- Source discipline. Scores are derived from named, curated open sources; per-headline business-impact briefs link to the underlying reporting.
- Calibration over drama. The rubric is conservative by instruction — a high score must be earned by evidence of operational disruption, not by headline volume.
- Corrections. Factual errors are corrected in the next publishing cycle once verified. Report suspected errors to info@warningofwar.com — corrections are version-tracked through the public git history.
What this is not
Warning of War is an information product, not professional advice. The scores and briefs are not investment advice, trading signals, legal or sanctions-compliance counsel, insurance guidance, or operational routing instructions. They are a structured second opinion intended to complement — never replace — professional risk management, qualified counsel, and your own judgement. Decisions taken on the basis of this site are the reader's own responsibility.
Find us elsewhere
- X (Twitter) — @Warning_of_War
- Substack — @warningofwar
- LinkedIn — company/warningofwar
For inquiries, corrections, press, or data licensing, see the contact page.