vinkius-labs / vurb.ts

**The MVA framework for production MCP servers.** Structured perception for AI agents. Zero hallucination. Zero data leaks. Documentation · Quick Start · API Reference --- Get Started in 5 Seconds That's it. A production-ready MCP server with file-based routing, Presenters, middleware, tests, and pre-configured connections for **Cursor**, **Claude Desktop**, **Claude Code**, **Windsurf**, **Cline**, and **VS Code + GitHub Copilot**. Choose a vector to scaffold exactly the project you need: | Vector | What it scaffolds | |---|---| | | file-based routing. Zero external deps | | | Prisma schema + CRUD tools with field-level security | | | n8n workflow bridge — auto-discover webhooks as tools | | | OpenAPI 3.x / Swagger 2.0 → full MVA tool generation | | | RFC 8628 Device Flow authentication | Drop a file in , restart — it's a live MCP tool. No central import file, no merge conflicts: --- Why Vurb.ts Exists Every raw MCP server does the same thing: the database result and ship it to the LLM. Three catastrophic consequences: **Data exfiltration.** sends , , — every column — straight to the LLM provider. One field = one GDPR violation. **Token explosion.** Every tool schema is sent on every turn, even when irrelevant. System prompt rules for every domain entity are sent globally, bloating context with wasted tokens. **Context DDoS.** An unbounded can dump thousands of rows into the context window. The LLM hallucinates. Your API bill explodes. --- The MVA Solution Vurb.ts replaces with a **Presenter** — a deterministic perception layer that controls exactly what the agent sees, knows, and can do next. The result is not JSON — it's a **Perception Package**: No guessing. Undeclared fields rejected. Domain rules travel with data — not in the system prompt. Next actions computed from data state. --- Before vs. After **Before** — raw MCP: **After** — Vurb.ts with MVA: The handler returns raw data. The Presenter shapes absolutely everything the agent perceives. --- Architecture Egress Firewall — Schema as Security Boundary The Presenter's Zod schema acts as a whitelist. **Only declared fields pass through.** A database migration that adds doesn't change what the agent sees — the new column is invisible unless you explicitly declare it in the schema. DLP Compliance Engine — PII Redaction GDPR / LGPD / HIPAA compliance built into the framework. compiles a V8-optimized redaction function via that masks sensitive fields **after** UI blocks and rules have been computed (Late Guillotine Pattern) — the LLM receives instead of real values. Custom censors, wildcard paths ( , ), and centralized PII field lists. **Zero-leak guarantee** — the developer cannot accidentally bypass redaction. 8 Anti-Hallucination Mechanisms Each mechanism compounds. Fewer tokens in context, fewer requests per task, less hallucination, lower cost. FSM State Gate — Temporal Anti-Hallucination **The first framework where it is physically impossible for an AI to execute tools out of order.** LLMs are chaotic — even with HATEOAS suggestions, a model can ignore them and call with an empty cart. The FSM State Gate makes temporal hallucination structurally impossible: if the workflow state is , the tool **doesn't exist** in . The LLM literally cannot call it. | State | Visible Tools | |---|---| | | , | | | , , | | | , | | | | Three complementary layers: **Format** (Zod validates shape), **Guidance** (HATEOAS suggests the next tool), **Gate** (FSM physically removes wrong tools). XState v5 powered, serverless-ready with . Zero-Trust Sandbox — Computation Delegation The LLM sends JavaScript logic to your data instead of shipping data to the LLM. Code runs inside a sealed V8 isolate — **zero access** to , , , , , . Timeout kill, memory cap, output limit, automatic isolate recovery, and AbortSignal kill-switch (Connection Watchdog). auto-injects HATEOAS instructions into the tool description — the LLM knows exactly how to format its code. Prototype pollution contained. escape blocked. One isolate per engine, new pristine context per call. State Sync — Temporal Awareness for Agents LLMs have no sense of time. After then , the agent still believes the list is unchanged. Vurb.ts injects RFC 7234-inspired cache-control signals: Registry-level policies with , glob patterns ( , ), policy overlap detection, observability hooks, and MCP emission. Prompt Engine — Server-Side Templates MCP Prompts as executable server-side templates with the same Fluent API as tools. Middleware, hydration timeout, schema-informed coercion, interceptors, multi-modal messages, and the Presenter bridge: decomposes any Presenter into prompt messages — same schema, same rules, same affordances in both tools and prompts. Multi-modal with , , . Interceptors inject compliance footers after every handler. with filtering, pagination, and lifecycle sync. Agent Skills — Progressive Instruction Distribution **No other MCP framework has this.** Distribute domain expertise to AI agents on demand via MCP. Three-layer progressive disclosure — the agent searches a lightweight index, loads only the relevant SKILL.md, and reads auxiliary files on demand. Zero context window waste. Skills follow the agentskills.io open standard — SKILL.md with YAML frontmatter. returns the lightweight index. returns full instructions. gives access to auxiliary files with **path traversal protection** (only files within the skill's directory). Custom search engines supported. Fluent API — Semantic Verbs & Chainable Builders Every builder method is chainable and fully typed. Types accumulate as you chain — the final has 100% accurate autocomplete with zero annotations: embeds prompt engineering. prevents backend overload. caps response size. streams MCP progress notifications. / / control temporal awareness. enables computation delegation. enables FSM gating. Middleware — Pre-Compiled, Zero-Allocation tRPC-style context derivation. Middleware chains com…