Infrastructure and Developer Surfaces

Scope

Packaging, deployment, developer tooling, testing infrastructure, and governance for the Hyperon ecosystem. This family covers the surface area that developers and operators interact with — how to install, test, distribute, and contribute to the project. Some repos here also appear in MeTTa Implementations where they serve a dual role as both tooling and implementation ecosystem components.

Active Repositories

Hyperon Developer Tooling

Repo	Language	Upstream	Maturity	Purpose
metta-lsp	TypeScript	iCog-Labs-Dev	Operational	VS Code LSP extension: Tree-sitter parsing, diagnostics, go-to-definition, hover, completion, rename, formatting. (Also in MeTTa Implementations.)
jupyter-petta-kernel	Python	trueagi-io	Operational	Jupyter kernel for MeTTa execution via PeTTa. Requires PETTA_PATH + SWI-Prolog + janus. (Also in MeTTa Implementations.)
metta-prebuilt-binary	Rust / Shell	iCog-Labs-Dev	Operational	Prebuilt MeTTa binary distribution with metta-run CLI formatter (v0.1.11). curl install script. (Also in MeTTa Implementations.)
metta-catalog	JSON	trueagi-io	Operational	Official JSON registry of public MeTTa modules for package discovery. (Also in MeTTa Implementations.)
metta-testsuite	MeTTa / Shell / Python	logicmoo	Operational	Cross-interpreter compatibility test suite with CI, JUnit conversion, and nightly reports. (Also in MeTTa Implementations.)
MettaWamJam	SWI-Prolog / Python	patham9	Operational	HTTP REST server exposing MeTTa execution via PeTTa. Docker-ready. (Also in MeTTa Implementations.)
zed-metta-mcp	Rust (planned)	JungeWerther	Pre-development	MCP server for MeTTa integration. Skeletal — README + LICENSE only, no code yet.
magi-playground-wiki	Python	—	Sandbox scaffold	Docker-isolated MeTTa evaluation sandbox (--network=none, read-only, --cap-drop=ALL, 256 MB, 5 s eval / 10 s wall) for the wiki.hyperon.dev interactive MeTTa playground surface
hyperon-wiki-mcp	Ruby	—	Implementation-backed	MCP server exposing an eight-tool AtomSpace mirror read surface (query_atoms, get_card_atom, get_card_provenance, list_references, list_atoms_by_type, atom_types, atom_count_by_type, space_stats) gated by mcp:atomspace:read scope

Legacy Packaging and Deployment

Repo	Language	Upstream	Maturity	Purpose
docker	Dockerfile / Shell	opencog	Operational (core); Legacy (ROS)	Docker images for classical AtomSpace, CogServer, language learning, and tutorials. ROS/Eva demo images are bitrotted.
ocpkg	Shell	opencog	Operational	Shell scripts (octool) automating classical-stack dependency install, build, and test on Ubuntu 22.04+.
opencog-debian	Shell / Debian	opencog	Operational	Debian/Ubuntu .deb packaging for 9 classical-stack components. Maintained by mhatta. Packages on Debian sid.
guix-atomese	Guile Scheme	opencog	Experimental	GNU Guix channel packaging Atomese ecosystem (cogutil through link-grammar). Newest packaging approach; container mode broken.

Legacy Infrastructure

Repo	Language	Upstream	Maturity	Purpose
cogutil	C++	opencog	Independently maintained	Foundational C++ utility library (v2.2.1, C++20). Threading, logging, signals. Required by all classical C++ repos. Independently maintained by the original classical-stack contributors.
cogserver	C++	opencog	Independently maintained	Network server for AtomSpace (v3.4.1). Telnet, HTTP, WebSocket, JSON, MCP interfaces. Independently maintained. (Also in Knowledge Substrates.)
benchmark	C++ / Python	opencog	Operational (partial)	Benchmarking tools for AtomSpace and CogServer performance. query-link and query-loop work; micro/ subdirectory broken.
logicmoo_cogserver	SWI-Prolog	opencog	Operational	CogServer emulator in SWI-Prolog for LogicMOO/classical-stack interop. Bridges Prolog and AtomSpace worlds. Default port 12101.

Governance

Repo	Language	Upstream	Maturity	Purpose
sneps	Markdown	singnet	Operational	SingularityNet Enhancement Proposals — governance process for ecosystem-wide standards (4 proposals).

How They Fit Together

Hyperon developer workflow: metta-lsp provides IDE support → metta-testsuite validates cross-implementation compatibility → metta-prebuilt-binary distributes binaries → metta-catalog resolves modules → jupyter-petta-kernel enables notebook exploration → MettaWamJam provides web API access.

Legacy packaging: ocpkg automates source builds on Ubuntu → opencog-debian produces .deb packages → docker provides containerized builds → guix-atomese offers reproducible Guix builds (newest approach, still experimental).

Legacy infrastructure: cogutil is the root dependency → cogserver provides network access → benchmark validates performance → logicmoo_cogserver bridges Prolog.

Quick Start

# Install MeTTa (prebuilt binary)
curl -sSL https://raw.githubusercontent.com/iCog-Labs-Dev/metta-prebuilt-binary/main/install.sh | bash

# VS Code extension
cd metta-lsp && npm install && npm run build  # or install from marketplace

# Jupyter kernel
cd jupyter-petta-kernel && ./install.sh  # requires PETTA_PATH set

# Legacy classical stack (Ubuntu, automated)
cd ocpkg && ./ocpkg -rdpcav -l default

# Docker (classical stack)
cd docker/opencog && ./docker-build.sh -a

Excluded from This Family

• opencog-nix: Historical Nix packaging effort pinned to old Python 3.6. Not actively maintained.
• opencog-tutorials, opencog-workshops: Legacy educational content. Covered in History.
• mcp-xp: Galaxy bioinformatics MCP chatbot — placed in Neural-Symbolic and LLM Integration.

Gaps and Consolidation Opportunities

• No unified Hyperon installer: hyperon-experimental installs via pip; PeTTa via shell scripts; MORK via cargo nightly. No single command installs a working Hyperon stack.
• Packaging split across 4 systems: ocpkg (Ubuntu scripts), opencog-debian (.deb), guix-atomese (Guix), docker (containers). Each covers different components at different freshness levels.
• zed-metta-mcp has no code: Pre-development only. Track for future MCP server development.
• metta-testsuite is not enforced in CI: Implementations don't consistently run it. Making it a required CI step would improve cross-implementation convergence.