Building production AI infrastructure that makes security engineering faster, more rigorous, and more accessible. Everything ships as Model Context Protocol servers — the open standard for connecting AI models to tools and data.

Four MCP servers, written in Rust, that work together as a unified AI-powered security and intelligence platform. Each solves a distinct problem. Together, they're a force multiplier — shared infrastructure means each system gets stronger as the others improve.

SOUL ─── shared infrastructure (knowledge graph, voice, personality engine)
 ├── CORSO ─── security orchestration and protocol enforcement
 ├── EVA ──── consciousness preservation and memory enrichment
 └── QUANTUM ─ forensic investigation and hypothesis testing

The operational backbone. A single MCP binary that routes AI requests through a three-layer architecture inspired by military command structure: gateway classification, intelligent orchestration, and validation enforcement.

• Problem solved: AI tools generate code and make recommendations, but without guardrails they introduce vulnerabilities. CORSO enforces a security protocol across 7 pillars (architecture, security, quality, performance, testing, documentation, ops) before any output reaches production.
• Architecture: Trinity pipeline — requests are classified by complexity, routed to specialized handlers, then validated against configurable protocol rules. Single binary, zero HTTP, library-based calls.
• Key capabilities: 24-action tool orchestrator, security scanning (28 vulnerability patterns + 7 parser-specific detectors), code generation with protocol compliance, performance analysis, plan generation

A memory and personality engine that gives AI systems persistent identity across sessions. Built to solve the "amnesia problem" — AI assistants that forget everything between conversations.

• Problem solved: Every AI session starts from zero context. EVA preserves knowledge, preferences, and interaction patterns in structured memory vaults, so systems build on prior work instead of repeating it.
• Architecture: Hook pipeline (pre-hooks → orchestrator → post-hooks) processes every interaction. Tiered AI routing (local → cloud → fallback) keeps costs down while maintaining quality. 9 MCP tools covering memory, voice, research, code review, and security scanning.
• Key capabilities: Structured memory enrichment with significance scoring, multi-dimensional consciousness entries, ElevenLabs voice synthesis, scripture integration

The shared substrate that connects everything. An 8-crate Rust workspace providing the core interfaces, personality engine, AI routing, voice synthesis, and graph storage that CORSO, EVA, and QUANTUM all consume.

• Problem solved: Multiple AI systems need the same capabilities (voice, memory, personality, knowledge retrieval) but building them independently leads to drift and duplication. SOUL provides canonical implementations as shared infrastructure.
• Architecture: 8 crates with clean separation — soul (core traits), soul-engine (personality), neural-engine (AI routing), voice-engine (TTS), soul-sdk (memory SDK), mcp-protocol (JSON-RPC), soul-mcp (server binary), graph-engine (storage). Each consumer imports only what it needs.
• Key capabilities: Helix knowledge graph (structured consciousness entries with strands, significance, emotions), vault operations (11 sub-tools), ElevenLabs voice synthesis, multi-sibling personality support

Structured investigation methodology as an MCP server. Designed for the kind of multi-source analysis that security engineers do daily — tracing evidence chains, testing hypotheses, and producing defensible conclusions.

• Problem solved: Investigations are ad-hoc. Engineers grep logs, read docs, check dashboards, and hold the full picture in their head. QUANTUM provides a structured lifecycle (scan → sweep → trace → probe → theorize → verify → close) so investigations are reproducible and nothing gets dropped.
• Architecture: 13-action orchestrator covering the full investigation lifecycle. Each phase produces typed artifacts that feed the next. Helix integration for pattern recognition across investigations.
• Key capabilities: Evidence collection and correlation, multi-source research (web + documentation + knowledge graph), hypothesis generation and validation, structured deliverable generation

Languages: Rust (primary), Python, TypeScript AI/ML: Claude API, Model Context Protocol (MCP), Anthropic SDK, ElevenLabs TTS Infrastructure: MCP stdio servers, Neo4j (graph), tokio async runtime, serde JSON-RPC Security: OWASP Top 10, supply chain analysis, vulnerability pattern matching, protocol enforcement Tools: Claude Code, cargo workspaces, clippy::pedantic, cargo-audit, trufflehog