msbardelottoribeiro-blip/project-argus

Project Argus is an early-stage, open source project that wants to build an automated watchdog for United Nations spending. The idea is to create a system that can look at UN procurement contracts and humanitarian aid funds and flag things that look wrong, such as overpricing, unusual patterns, or signs of possible corruption, ideally before the money is actually spent. The plan is for this to work through several AI agents working together rather than one single program. These agents would continuously scan public UN financial data, cross check details against each other, and surface anomalies for humans to review. The end goal described by the project is a public dashboard anyone can check, so the process of catching fraud or waste in international aid becomes transparent rather than hidden inside internal audits. Right now the project is only proposing the tools it wants to use, not yet using them fully. The plan mentions Python with libraries like Scrapy and BeautifulSoup for pulling in data from external sources, Neo4j as a graph database for mapping relationships between shell companies and politically exposed persons, a Llama 3 based AI pipeline for reading documents in multiple languages, and a React or Next.js frontend for the dashboard itself. The project describes itself as being in Phase 1, which is about building the data architecture. The README asks specifically for Python developers to help create the first scrapers that pull data from the UN Global Marketplace and UN OCHA APIs. There is no working demo, no live dashboard, and no finished audit pipeline yet: this is a call for early contributors to help lay the groundwork. If you are non technical, think of this as a proposal for a public fraud detection tool for UN spending, still in its planning and setup stage, looking for coders to help build the first data collecting pieces.