Hermes Platform Architecture
1. Observe What changed? β
2. Understand Why does it matter? β
3. Track How does risk evolve? β
4. Predict What happens next? β
5. Decide What to do? β
6. Verify What happened? β
7. Remember Persistent memory
ποΈ Platform Pillars
Section titled βποΈ Platform Pillarsβ π‘ Machine-Readable Data & Feeds Raw JSON APIs, OASIS STIX 2.1 bundles, and TAXII 2.1 discovery collections.
πΈοΈ Knowledge Graph Ontology Graph database connecting 556+ entities and 575+ relationships across threats.
βοΈ Epistemic Methodology Mathematical foundations for HTS, risk curves R(t), and confidence vectors.
π‘οΈ Trust & Security Center 100% client-side privacy model, reproducible builds, and data limitations.
1. Zero-Backend Sovereign Architecture
Section titled β1. Zero-Backend Sovereign ArchitectureβUnlike legacy cybersecurity SaaS platforms that centralize and store proprietary client inventories in remote clouds:
- Client-Side Execution: The Hermes UI, SBOM analyzer, and risk comparator execute entirely within your browser runtime using WebAssembly and modern JavaScript.
- Static Pre-Compilation: All threat intelligence feeds and graph indexes are pre-compiled during deterministic Linux build stages.
- No Central Ingestion of Your Data: Your SBOMs, CVE matches, and configuration files never leave your workstation.
2. Ingestion & Intelligence Factory Pipeline
Section titled β2. Ingestion & Intelligence Factory PipelineβINGEST (NVD, CISA KEV, EPSS, arXiv, Git) βNORMALIZE (Canonical Schema Validation) βOBSERVE (Epistemic State & Confidence Vector OBS-*) βGRAPH (Entity Ingestion & Relationship Matrix) βTRAJECTORY (Derive R(t) & Discrete Velocity) βFORECAST (Probabilistic Bounded Predictions) βDECIDE (Prescriptive T0/T1 Directives) βPUBLISH (Static JSON, STIX 2.1, TAXII 2.1, HTML)