omega-mcp
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
omega-mcp
An MCP server, written in Omega, that lets an agent compile and run Omega (.om) programs with the
pinned seeded compiler and look up the language. It speaks MCP (protocol version 2025-06-18) as
newline-delimited JSON-RPC 2.0 over stdio.
Pin: omega-mv e5b32567a7aff3ce594e3254a622a846a6cd190b, omc sha256
2c82480596bbdc50cda9626866209ef662d5834f278f839ff4ce0250c495029f. build.sh refuses any other
compiler, and the server checks the sha256 of its compiler again at start-up.
Tools
Tool | Arguments | Result |
|
|
|
|
| compile result, capped stdout/stderr, and the exit: |
|
| meaning, cause, fix, a minimal example and citations into omega-mv@e5b32567 |
|
| the entry (form, arity, types, result, example, citations), or up to 20 search hits |
A compile diagnostic is a successful tool result (isError: false); isError is true when nothing
could be run (sandbox failure, a refused compile stage) or when explain/lookup found no entry. Bad
arguments are JSON-RPC errors (-32602).
Related MCP server: Ephemora Cell MCP
Isolation
Each compile and run stage gets a fresh /tmp/omega-mcp-<16 hex> directory (mode 0700, removed
after the reply) and runs under bwrap --unshare-all --clearenv. The sandbox sees only the compiler
or the program (read-only), its in-sandbox reaper and a 16 MiB tmpfs working directory: no network,
no host files, an empty environment. prlimit applies the resource limits to the sandbox before it
starts, and the server kills the whole sandbox at the wall clock.
Stage | CPU | Address space | File size | Processes | Wall clock |
run | 10 s | 256 MiB | 1 MiB | 5 |
|
compile | 20 s | 512 MiB | 16 MiB | 4 | 20 s |
Both stages also get 64 open files and no core dumps. --test-only-no-isolation turns all of this
off and exists only so the trust gate can prove it detects an escape; never give it to an agent.
Build and use
OMEGA_SRC=/path/to/omega-mv ./build.sh # omega-mv checkout at e5b32567 with seed/out/omc built
build/omega-mcp # reads JSON-RPC on stdin; logs one line per request on stderrThe server loads data/reference.json and data/examples/ from beside its executable. Start-up
exit codes: 2 bad arguments, 3 compiler not the pin, 4 data missing, 5 bwrap/prlimit missing.
mcp.json is an example client configuration; point its command at your build/omega-mcp.
Gates
python3 gates/smoke.py # MCP client session; transcript in evidence/smoke-transcript.jsonl
python3 gates/trust.py # trust boundary: 4 attacks plus negative controls; evidence/trust-gate.json
python3 tools/check-reference.py --omega-src $OMEGA_SRC --omc $OMEGA_SRC/seed/out/omc --self-testtrust.py sends, through the run tool, a read of a canary under ~/.local/state, a fork bomb, an
infinite loop and a 10 GiB write, and requires each to be refused or killed with no process left
behind. A second server started with --test-only-no-isolation runs bounded versions of the same
attacks and must see them succeed, so a gate that cannot see an escape fails. check-reference.py
re-derives every keyword, error and trap entry from the pinned source, compiles and runs every example
with the pinned omc, and runs three negative controls.
These gates are host evidence on Linux arm64 (bubblewrap 0.9.0, util-linux 2.39.3); they say nothing
about other kernels or about a sandbox configuration other than the one in src/sandbox.om.
Layout
src/*.om: the server (src/UNITSlists the concatenation order;build.shbuilds it)vendor/omega-mv/: sha256 and JSON units copied unchanged from omega-mv@e5b32567data/: reference data and example programs, written bytools/gen-reference.pyand checked bytools/check-reference.pygates/,evidence/: verification oracles and their recorded resultsDEPENDENCIES.md: every host tool in the build, runtime and verification paths
License: Apache-2.0 (see LICENSE and NOTICE).
Human-Directed Agentic Engineering
Development follows Human-Directed Agentic Engineering (HDAE): humans define goals, architecture, constraints, acceptance criteria, and stopping points; agents implement, execute checks, diagnose failures, and revise within those boundaries. Human review retains responsibility for claims, acceptance, and release decisions. This describes the development method, not a change to runtime autonomy or historical evidence.
For the documented method and a systems-software case study, see Gregory M. Kurtzer (2026), Ginux: Building an Independent Kernel That Runs Linux Userspace through Human-Directed Agentic Engineering, technical report, v1, Zenodo, https://doi.org/10.5281/zenodo.23682548.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
Secure Linux sandboxes for AI agents, as MCP tools.
Hosted one-shot code execution: send a snippet plus optional files, get stdout, stderr, exit...
MCP server for Superserve sandboxes: create, exec, and manage Firecracker microVMs
Hosted MCP runtime where your agent publishes its own tools by tool call and keeps working.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceMCP server that executes caller-supplied Python code in a network-isolated, locked-down sandbox and returns stdout/stderr/exit code. Supports stdio, HTTP, and SSE transports with optional auth.1MIT
- AlicenseAqualityAmaintenanceDependency-free, locally sandboxed MCP stdio server. Tools run as WASM inside a capability-based WASI sandbox (wasmtime); every tools/call is attested with cost/policy and signed manifests fail-closed.367 PyPI48Business Source 1.1
- FlicenseNot gradedqualityBmaintenanceEnables agents and MCP clients to safely execute and verify multi-language code in a deterministic sandbox, using AST security analysis to block dangerous system calls and capture isolated execution telemetry.7-
- FlicenseNot gradedqualityNot gradedmaintenanceEnables agents and MCP-compliant clients to run multi-language code in a deterministic, zero-dependency sandbox where AST security analysis blocks dangerous syscalls and isolated execution telemetry is captured. It exposes this through JSON-RPC 2.0 stdio, letting frameworks like LangGraph, CrewAI, Cursor, and Claude Desktop validate and execute agent-generated code safely.-