omni-mcp-gateway
Rust Enterprise MCP Fleet, Tool Groups & Code Mode Gateway v6.0 (macOS, Linux, Windows x64/ARM64): rustls TLS/mTLS, OIDC/JWKS (RS256/ES256/EdDSA), tokio async, SQLite WAL+CAS, wasmi Wasm sandbox, Okapi BM25 0-turn routing, SSA parallelization, Saga rollbacks & Fleet UI.
Install / Use
claude mcp add Rangow4562 -- npx -y github:Rangow4562/omni-mcp-gatewayIf the server publishes to npm under a different name, use that package instead — check the repo README.
MCP Server
Model Context Protocol server
Quality Score
Category
AI & Machine LearningSupported Platforms
Our assessment of omni-mcp-gateway
omni-mcp-gateway scores 74/100 on our quality scale, 853rd of 957 AI & Machine Learning skills we index.
Its MCP Server is 40 KB long, well organised into 47 sections with 13 code examples: long enough that it reads more like full documentation than a focused instruction file, which agents can find harder to follow.
It has 3 GitHub stars, so there is little community track record yet; judge it on its content.
Maintenance, license and trust
- The repository was last updated yesterday, so omni-mcp-gateway is actively maintained.
- No license is declared. By default that means all rights are reserved: you can read it, but reusing or redistributing it is not clearly permitted. Ask the author before building on it commercially.
- Its trust signals score 80/100, with 2 cautions from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.
Safety scan
No issues foundOur scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.
AI review by kimi-k2.7-code on 2026-11-16. Automated pattern scan on 2026-11-16. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.
omni-mcp-gateway compared with similar skills
All 4 of these similar skills score higher than omni-mcp-gateway; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| omni-mcp-gateway (this skill)by Rangow4562 | 74 | 3 | 1d ago | MCP Server |
| claude-memby thedotmack | 100 | 117.7k | today | CLAUDE.md |
| Agent-Reachby Panniantong | 100 | 112.4k | 9d ago | CLAUDE.md |
| Understand-Anythingby Egonex-AI | 100 | 92.9k | 1d ago | CLAUDE.md |
| rufloby ruvnet | 100 | 81.9k | today | CLAUDE.md |
Frequently asked questions
- How do I install omni-mcp-gateway?
- Run
claude mcp add Rangow4562 -- npx -y github:Rangow4562/omni-mcp-gateway. The install tabs above show the steps for each supported agent. - Which AI agents does omni-mcp-gateway work with?
- It is written for Claude Code and Claude Desktop, as a MCP Server file. Other agents that read the same format can often use it too.
- Is omni-mcp-gateway safe to use?
- Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It declares no license and scores 80/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is omni-mcp-gateway still maintained?
- The repository was last updated yesterday, so omni-mcp-gateway is actively maintained.
Skill content
View source on GitHubOmni MCP (omni-mcp v5.5 — Autonomous Runtime & Enterprise MCP Gateway)
A high-performance, zero-dependency Rust enterprise Model Context Protocol (MCP) management platform and runtime gateway combining native in-process Omni Code Mode with enterprise service mesh governance, Okapi BM25 intent ranking, columnar predicate pushdown, resumable continuation snapshots, and declarative Saga compensations.
Omni MCP unifies hundreds of microservices and thousands of tools behind a standardized, sub-millisecond, token-optimized Rust gateway with:
- Native Code Mode Execution (
codemode): In-process sandbox executing JavaScript/TypeScript syntax scripts with closure syntax (|x| expr) and service host calls in $\sim 0.10\text{ ms}$ ($100\ \mu\text{s}$). - Resumable AST Continuations: Stateless suspend/resume of execution state (
SuspendedExecutionFrame) across approval gates (is_mutation: true) via durable FNV-1a checksummed WAL snapshots. - Zero-Copy Columnar Predicate Pushdown: Compiles
.filter()and.map()AST closures directly into vectorizedPhysicalExprpipelines over contiguous columnar batches (ColumnarRecordBatch). - Static AST SSA Auto-Parallelization: Automatic dependency analysis lifts independent sequential
await codemode.*calls into concurrentStmt::ParallelWavebatches without requiring complex LLMPromise.allemissions. - Declarative Saga Compensation Rollbacks: Ephemeral execution transaction logs unwind completed mutations in reverse (
LIFO) order when a downstream step fails. - Speculative 0-Turn Schema Pre-Injection: True Okapi BM25 ($k_1 = 1.2, b = 0.75$) + stem routing injects the top-$K$ highest-confidence
.d.tsdeclarations on Turn 1 based on natural-language user prompts. - Deterministic Simulation Testing (
dst) Engine: 64-bitXoshiro256StarStarseeded PRNG fault-injection suite validating agent scripts against network drops, circuit trips, latency spikes, and malformed payloads. - Authoritative
-320xxError Contract & Multi-Revision MCP Protocol Negotiation: Compiler-checkedFeatureStatusenum mapping every runtime guardrail (-32030..-32038) and negotiating official MCP revisions (2024-11-05,2025-03-26,2025-06-18,2025-11-25), with graceful fallback for future spec revisions like2026-07-28and statelessserver/discoverover Streamable HTTP. - Live HTTP/1.1 Downstream Transport (
transport-http): Feature-gated, pure standard library connection-pooled HTTP client with timeouts, keep-alive, and jittered exponential retries feeding per-service circuit breakers and Saga rollbacks. - Schema Preflight Validator (
-32035 SCHEMA_PREFLIGHT_FAILED): Static AST input schema type-checker detecting all violations in a single pass before execution or approval gates for single-turn LLM self-correction. - HMAC-SHA256 Authenticated Continuations & Approval Hardening: Tamper-proof continuation frames signed via HMAC-SHA256 with atomic compare-and-swap (CAS) single-use binding (
-32038 APPROVAL_INVALID), approver identity and argument digest verification, and crash-safe WAL/snapshot Saga compensation ledger. - Comprehensive Sandbox Resource Limits: Strict execution guardrails bounding instruction fuel (
-32032), string/array heap growth (-32036 MEMORY_EXHAUSTED), wall-clock time (-32037 TIMEOUT), and internal panic containment (catch_unwind). - Dynamic TypeScript Declarations (
declare namespace codemode): Automatically converts OpenAPI 3.1, Protobuf, and Thrift JSON Schemas into compact.d.tsdeclarations, compressing tool prompt token footprints by $65%$ to $95%$. - Search & Execute Pattern: Built-in
codemode.search(query)andcodemode.describe(server, tool?)allowing agents to discover and inspect typed definitions on demand without context bloat. - Zero-Outbound Isolation & Approval Gates:
globalOutbound: nullisolation blocks rawfetch()or filesystem calls, while human-in-the-loop approval gates intercept state-mutating operations (is_mutation: true). - Full Mesh Governance: Wire envelope framing & protocol translation (gRPC, HTTP, simulated RPC), circuit breaking, semantic caching, ABAC security, OpenTelemetry W3C distributed tracing, canary weighted routing, and SHA-256 Merkle audit trail.
🛡️ What's New in Version 5.0 ("Trust Boundary Reset & MCP 2026-07-28 Conformance")
- 🤝 Full MCP
2026-07-28Protocol Conformance (B1–B10):- Required
"resultType": "complete"on all results and"resultType": "input_required"withrequestStatefor multi-round-trip requests (B1). - Deterministic lexicographical ordering and caching metadata (
ttlMs: 300000,cacheScope: "workspace") acrosstools/list,resources/list,prompts/list, andresources/templates/list(B2). - Streamable HTTP protocol header validation on POST
/omni/mcpand/<service>/mcp: requiresMcp-MethodandMcp-Name, rejecting mismatches with HTTP 400 and-32020 HEADER_MISMATCH(B3). - Pruned deprecated endpoints (
ping,logging/setLevel,roots/list_changed) and validated client capabilities with-32021 MISSING_REQUIRED_CLIENT_CAPABILITY(B4). - Spec-reserved error code allocation (
-32020..-32099for spec,-32000..-32019for gateway) with backward-compatibledata.legacy_codemappings (B5). - RFC 9728 protected resource metadata at
/.well-known/oauth-protected-resource, HTTP 401WWW-Authenticate: Bearer resource_metadata=..., and HTTP 403insufficient_scope(B6). - MCP Tasks extension (
io.modelcontextprotocol/tasks) exposingtasks/getandtasks/updatefor suspended execution frame lifecycle management (B7). - Deprecated legacy
/sseendpoint returning410 Goneby default (opt-in--features legacy-sse) (B8).
- Required
- 🛡️ Universal Route Authorization & Atomic Approvals (A1–A6):
- Deny-by-default role-based access control (RBAC) across all
/api/*and/api/v1/*endpoints (Agent,Operator,Approver,Admin). - Atomic compare-and-swap (CAS) frame consumption under write lock preventing race conditions on multi-approver resumes.
- Separation of duties cryptographically forbidding mutation requesters from approving their own actions.
- Length-prefixed HMAC-SHA256 continuation signing over all 12 frame fields with argument digest verification.
- Deny-by-default role-based access control (RBAC) across all
- 🌐 HTTP Front-Door Hardening & DoS Containment (A7–A10, A14):
- Recursive JSON parser depth guard (
MAX_JSON_DEPTH = 64) preventing stack-overflow crashes. - Bounded worker channel shedding saturated load with
503 Service Unavailable, plus connection-levelcatch_unwindpreventing thread death. - Persistent
BufReaderframing across HTTP/1.1 keep-alive loops with strict digit-onlyContent-Lengthchecks.
- Recursive JSON parser depth guard (
- ⚡ Architecture, Observability & Tooling (D7, D8, C1):
- $O(1)$ ring-buffer trace eviction via
VecDeque<TraceSpan>. - Versioned
/api/v1/*REST routes and OpenAPI 3.1 specification at/openapi.json. - CLI subcommand direct execution (
omni-gateway list,omni migrate).
- $O(1)$ ring-buffer trace eviction via
🤝 What's New in Version 4.7 ("Protocol Parity & Real Wire")
- 🤝 MCP Protocol Parity (
2024-11-05,2025-03-26,2025-06-18,2025-11-25) (M1): First-class protocol capability matrix supporting four active revisions with graceful fallback for2026-07-28. Implements statelessserver/discover,subscriptions/listen,_metaprotocol version and client capability negotiation, batch gating, W3C trace context bridging, and Streamable HTTP on/omni/mcpand/<service>/mcpwith deprecated legacy/sse. - 🌐 Live HTTP/1.1 Downstream Transport (
transport-http) (M2): Feature-gated, pure standard library HTTP/1.1 downstream client with connection pooling, keep-alive, connect/read/write timeouts, and bounded retries with jitter feeding the per-service circuit breaker and Saga unwinder. - 📋 Schema Preflight Validator (
-32035 SCHEMA_PREFLIGHT_FAILED) (M3): Static AST input schema type-checker validating parameters against registered schemas before approval gates or execution, collecting all violations in a single pass for single-turn LLM correction. - 🔐 Approval & Continuation Hardening + Persistent Saga Ledger (M4): Authenticated continuation frames via HMAC-SHA256, atomic compare-and-swap single-use approval binding (
-32038 APPROVAL_INVALID), OS entropy mixing, and crash-safe WAL/snapshot persistence of in-flight Saga compensations. - 🧱 Sandbox Resource Limits (M5): Comprehensive sandbox boundaries with memory limit tracking (
-32036 MEMORY_EXHAUSTED), wall-clock execution timeouts (-32037 TIMEOUT), and panic containment (catch_unwind). - 🛡️ HTTP Surface Hardening & Universal RBAC (H1): DNS rebinding protection via Origin validation, loopback bind by default with non-loopback startup guard, universal Bearer authentication and role-based access control (RBAC) across all
/api/*endpoints, slowloris timeouts, header line and size caps, and HTTP request smuggling protection.
🛡️ What's New in Version 4.6 ("Honesty & Hardening")
- 🧪 Code Mode Test Fortress (M1): In-module
#[cfg(test)]unit test suites acrosscodemode.rs,json.rs,wal_store.rs, andgateway.rsverifying parser diagnostics, evaluator closures, fuel exhaustion (-32032),globalOutbound: nullisolation (-32033), vectorizedPhysicalExprevaluation,SuspendedExecutionFrameround-trip resume identity, FNV-1a WAL corruption rejection, and Saga LIFO unwinding order. - 🎯 True Okapi BM25 Ranking (M2): Replaces additive substring scoring with full Okapi BM25 ($\text{IDF}$ corpus weighting, term-frequency saturation $k_1 = 1.2$, document length normalization $b = 0.75$, camelCase/snake_case splitting, suffix stemming, and field boosts) across
speculative_pre_inject,discover_server, andrecommend_tools. - 📐 Authoritative
-320xxError-Code Contract (M3): Strongly-typedFeatureStatusenum and structuredRPCErrorcontract covering-32030 CIRCUIT_OPEN,-32031 APPROVAL_REQUIRED,-32032 FUEL_EXHAUSTED,-32033 OUTBOUND_BLOCKED,-32034 SAGA_COMPENSATED, and-32035 SCHEMA_PREFLIGHT_FAILED. - 🤝 MCP Protocol Version Negotiation (M4): Lifecycle negotiation on
initializeechoing clientprotocolVersionwhen in{"2024-11-05", "2025-03-26"}(defaulting to"2025-03-26"), plus additive2025-03-26tool capability annotations (readOnlyHint,destructiveHint,idempotentHint,openWorldHint). - 🧹 Single-Sourced Versioning & Documentation Honesty: Rust runtime versions single-sourced from
env!("CARGO_PKG_VERSION")with an automated drift-guard test acrossCargo.toml,package.json,ui/package.json, andREADME.md.
🚀 What's New in Version 4.5
- 🔄 Resumable AST Continuations for Approval Gates (Stateless Suspend/Resume): Serializes instruction pointer, scope heap, and Merkle hashes into Crockford Base32 ULID (
SuspendedExecutionFrame) in durable WAL when hittingAPPROVAL_REQUIRED, resuming atinstruction_ptr + 1upon approval without re-running read tools. - ⚡ **Zero-Copy Columnar
Truncated for display — read the full file on GitHub.
Related Skills
claude-mem
117.7kPersistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More
Agent-Reach
112.4kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
Understand-Anything
92.9kGraphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more.
ruflo
81.9k🌊 The original agent harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, federation, vector RAG integration, and native Claude Code / Codex / Hermes and many more Integrated
Languages
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.
