SENTINEL-X
The Sixth Sense for Intelligence Operatives.
Created on 27th December 2025
•
SENTINEL-X
The Sixth Sense for Intelligence Operatives.
The problem SENTINEL-X solves
The Context: Intelligence operatives, government officials, and high-profile assets are increasingly becoming targets of coordinated disinformation campaigns. Hostile actors use "fake news," deepfakes, and social engineering to compromise agents' safety, leak their locations, or destroy their reputations.
The Gap: Current monitoring methods are:
Too Slow: Manual checking of news feeds cannot keep up with viral misinformation.
Too Noisy: It is impossible for humans to filter through thousands of "low-level" rumors to find the critical security threats.
Reactive: Agencies often find out about a narrative breach only after the damage is done.
How Sentinel X Solves It: Sentinel X acts as an Autonomous Digital Bodyguard. It is an AI-driven dashboard that ingests real-time global news data and applies intelligence logic to:
Detect Coordinated Attacks: It doesn't just look for keywords; it looks for patterns. A sudden spike in negative sentiment from a specific region triggers a high-level alert.
Quantify the Threat: Instead of vague warnings, it assigns a precise Threat Score (0-10) based on keyword severity (e.g., "leak," "assassinate"), source credibility, and velocity.
Visualize the Battlefield: The Global Threat Map instantly shows where the hostile narrative is originating, allowing for rapid geolocation of the threat source.
Real-World Application: This tool allows analysts to stop "doom-scrolling" and start decision-making. If a fake story about an operative surfaces in a specific city, Sentinel X flags it instantly, allowing the agency to initiate counter-measures or extract the asset before the situation escalates.
Challenges we ran into
- The "Demo Effect" & API Rate Limits The biggest hurdle was relying on live third-party APIs (NewsAPI) for a real-time dashboard. During testing, we realized that if the API rate limit was hit or the internet connection dropped during the demo, the entire dashboard would crash.
The Solution: I engineered a robust Hybrid Data Engine. I built a fail-safe system (the generate_sample_data function) that automatically detects API failures. If the live feed cuts out, the system seamlessly switches to high-fidelity synthetic data based on realistic threat patterns. This ensures the app never breaks during a presentation.
- Creating a Custom "Cyberpunk" UI in Streamlit Streamlit is great for data, but it notoriously forces a generic "white website" look. I wanted a specific "National Security/Dark Mode" aesthetic that felt like a movie interface.
The Solution: I had to dive deep into CSS Injection. By inspecting Streamlit's DOM elements, I wrote a custom MODERN_CSS module that overrides the default theming, adds neon glow effects to the metrics, and creates a responsive grid layout that native Streamlit doesn't support.
- Visualizing Heavy Geospatial Data Plotting hundreds of threat points on a map caused significant lag initially.
The Solution: I optimized the map rendering using Plotly Mapbox with a scattergl layer for performance. I also implemented a "jitter" algorithm to prevent markers from overlapping perfectly when multiple threats originate from the same city (e.g., "New York"), making the data readable without clutter.