
Overview
PROMETHEUS is an autonomous competitive decision agent built for enterprises that need more than a report — they need a verdict. The system runs four agents in sequence, each with a distinct adversarial role:
1. SCOUT AGENT — deploys Gemini 2.5 Flash with live Google Search grounding to build a real-time evidence base: market data, competitor filings, analyst reports, news signals. Outputs a structured brief with citations. 2. CHALLENGER AGENT — reads Scout's findings and is explicitly instructed to find every reason the opportunity is overrated, the data is misleading, or the thesis is flawed. Real devil's advocacy — not polite hedging. 3. STRATEGIST AGENT — reads both Scout and Challenger. Produces an integrated strategic brief that identifies the core disagreements, weights them, and constructs actionable scenarios. Returns a DisagreementMatrix. 4. DECISION GATE — synthesises all three prior agents and produces the final verdict: GO, NO-GO, or MONITOR, with a calibrated confidence score (0–100%), key conditions, and explicit rationale.
Three live scenarios are preloaded for instant demo: - OpenAI Market Entry → NO-GO at 65% confidence - Salesforce vs HubSpot Italy → MONITOR at 62% confidence - Tesla Competitive Threat Europe → GO at 88% confidence The UI populates each agent panel in sequence so judges can watch the reasoning unfold in real time — not just see the final answer. Tech stack: Gemini 2.5 Flash (all agents), Google Search grounding (Scout), FastAPI + Python backend deployed on Vultr Cloud Compute, Streamlit frontend on Streamlit Community Cloud, aiosqlite for analysis history.
PROMETHEUS does what no existing AI tool does: it ends with a decision.
Note: the demo URL is substituted with the GitHub repository link.
No comments yet
Be the first to comment on this project.

FloatGPT — Omnipresent, Local-First AI Desktop Companion
FloatGPT is an omnipresent desktop AI companion that floats over Windows to automate OS workflows, plan tasks, and generate real-time executive dossiers. Operating 100% privately on your PC or powered by the cloud, it transforms daily computing with screen awareness and calm execution.

ClaimTrace
ClaimTrace audits the citations in a scientific manuscript: it checks that every reference exists and is described correctly, and that each cited source actually says what the citing sentence claims, with a verbatim quote as evidence. On 1,728 reference test cases it never marked a fabricated or altered reference as verified.

seashield ai
Ocean buoys feed disaster-warning centres, which trust what they receive. A hacked buoy can fake giant waves to trigger a false alarm, or hide a real storm. SeaShield AI checks every reading with four layers: cryptographic signatures, physics rules learned from real data, an AI anomaly model, and comparison with nearby buoys. Suspicious alerts are held for human review. Our live demo on Vultr replays real NOAA data from Hurricane Helene: launch attacks yourself and watch SeaShield stop them.




