[![Project Logo](assets/logo-new.png)](https://aclas.college) # 🛡️ Aegis-Graph ### **The Sovereign Academic Audit Protocol** *Empowering Global Education with Sovereign Intelligence & Decentralized Graph Integrity*

Live Dashboard • Documentation • Whitepaper • Atlanta College of Liberal Arts and Sciences (ACLAS) Official Site

--- ## 🏛️ Executive Vision **Aegis-Graph** is a decentralized framework specifically engineered to safeguard academic integrity in the age of synthetic data. While technically governed by the **Atlanta College of Liberal Arts and Sciences (ACLAS)**, it operates as a global, sovereign protocol powered by **Agentic GraphRAG**. > "In an era of synthetic data, truth must be sovereign." — ACLAS Sovereign Node Group. --- ## ⚠️ Current Status The hosted dashboard is a **non-production UI demo**. It no longer issues browser-side credential approvals; professional verification requires server-side document parsing, issuer evidence, revocation checks, and a signed audit response. --- ## 🚀 Key Features

👁️ Vision Forensics

Planned pixel-level and OCR-based analysis to detect AI-generated artifacts in digital credentials and transcripts.

🗺️ Graph Navigator

Institution evidence resolution using a local index plus optional ROR lookup; ROR matches are supporting evidence, not automatic credential approval.

⚖️ Logic Auditor

Evidence-weighted consistency checks for registry status, timelines, blacklist aliases, and missing credential-authenticity proof.

🔒 Sovereign Ledger

Planned cryptographically signed audit trails for reproducible, tamper-evident credential review.

--- ## 🧠 The Agentic Architecture (MARS) Aegis-Graph operates via a collaborative **Multi-Agent Reasoning Swarm (MARS)**. Each agent handles a specific layer of the audit protocol, reaching consensus before issuing a final verdict. ```mermaid graph TD A[Digital Credential] --> B{MARS Swarm} B --> C[Vision Agent] B --> D[Graph Agent] B --> E[Logic Agent] C --> F[Pixel Forensics] D --> G[ROR Node Sync] E --> H[CoT Validation] F & G & H --> I[Consensus Layer] I --> J[Sovereign Audit Certificate] ``` --- ## 📊 Performance and Evaluation Production benchmark claims are not published yet. Current development focuses on eliminating browser-side approvals, normalizing graph evidence, and making audit decisions reproducible before publishing precision/latency metrics. * **Audit Precision**: pending reproducible benchmark suite * **Verification Latency**: pending server-side implementation * **Privacy Model**: local PII scrubbing prototype; ZK evidence is roadmap work * **Global Node Coverage**: local index plus optional ROR lookup --- ## 🌍 Global Localization (8-Language Matrix) Our dashboard and documentation are fully localized for international adoption: 🇺🇸 EN • 🇨🇳 CN • 🇪🇸 ES • 🇫🇷 FR • 🇩🇪 DE • 🇯🇵 JP • 🇰🇷 KR • 🇵🇹 PT --- ## 🛠️ Quick Start ```bash # Clone the Sovereign Registry git clone https://github.com/aclascollege/aegis-graph.git # Initialize the Multi-Agent Swarm pip install -r requirements.txt # Launch Local Auditor python main_pipeline.py ``` --- Aegis-Graph is governed by the **AEGIS-GRAPH Open Governance Board**, with technical support from the **ACLAS Sovereign Research Group**. We welcome contributions from researchers, developers, and academic institutions worldwide. - [Contributing Guidelines](CONTRIBUTING.md) - [Code of Conduct](CODE_OF_CONDUCT.md) - [Security Policy](SECURITY.md) ---