ellmos-ai stacks

🇩🇪 Deutsche Version

Note

ellmos-ai/stacks serves as the central schema specification and catalog index for all manifest-driven AI agent stacks within the ellmos-ai ecosystem. For machine-readable context, see llms.txt or docs/manifest-schema.md.

The catalog and composition-principle overview for every stack in the ellmos-ai ecosystem.

Machine-readable context for LLMs and agentic coding tools: llms.txt.

A "stack" here is a self-contained, publishable bundle: a manifest that lists which existing modules it composes and how they are wired together, plus (usually) a small installer. A stack does not duplicate the code of the modules it lists — it clones and connects them. This repository is the single place that documents the shared manifest schema and indexes every stack, so each stack's own README can stay focused on its own domain instead of re-explaining the composition principle.

Search and disambiguation

Use ellmos-ai/stacks or ellmos-ai stack catalog when looking for this repository. It is an ecosystem catalog and manifest-schema reference for AI-agent stacks; it is not the Stacks Bitcoin blockchain, a generic cloud or web-development stack, or a hosted LLM platform. The catalog's actionable entry points are the stack repositories it links, especially ellmos-ai/agent-ops-stack.

Architecture & Ecosystem Composition

graph TD
    Catalog["ellmos-ai/stacks (Catalog & Schema)"] --> ManifestSpec["Manifest Schema (ellmos-stack-manifest-v1)"]
    Catalog --> Stack1["ellmos-stack (Research & Knowledge Automation)"]
    Catalog --> Stack2["agent-ops-stack (Multi-Agent CLI Coordination)"]
    
    Stack1 --> M1["Ollama / Rinnsal / KnowledgeDigest / n8n"]
    Stack2 --> M2["FileCommander / CodeCommander / ControlCenter / LockMaster / SyncMaster"]

Loading

Start here

If you need...	Start with
A list of every stack and what it's for	Catalog below
To understand what a "stack" is and why manifests, not code copies	What is a stack below
The shared manifest field reference	docs/manifest-schema.md
A worked example of a manifest + installer	ellmos-ai/agent-ops-stack

Catalog

Stack	Focus	Status	Repository
ellmos-stack	Self-hosted AI research & knowledge automation (Ollama + n8n + Rinnsal + KnowledgeDigest + research pipeline)	Active	ellmos-ai/ellmos-stack
agent-ops-stack	Multi-agent coordination and personal-assistant tooling for local CLI coding agents: ticket routing, file locking, cross-machine sync, a decision-avatar, a shared skill format, and an MCP control plane	Active	ellmos-ai/agent-ops-stack
ellmos-research-stack	Academic research & literature: PubMed/arXiv pipelines, bibliography tools, citation networks	Planned	—
ellmos-dev-stack	Software development & DevOps: code analysis, CI/CD integration, repo monitoring	Planned	—
ellmos-media-stack	Content creation & media: transcription, summarization, media processing pipelines	Planned	—

The ellmos-stack row and the planned research/dev/media specializations originate from the Stack Family table in ellmos-ai/ellmos-stack; this catalog is the intended long-term single source for that family. agent-ops-stack is a second, independent stack family: it composes local agent-ops tooling (coordination between AI coding agents on one machine/user) rather than a server-side research automation stack.

What is a stack

The organizing idea, borrowed from the ellmos-sovereign composition principle: the installation is the blueprint. Instead of writing prose documentation that describes how components should be wired and hoping the real install matches it, a stack ships one manifest file that an installer reads and acts on directly. The manifest is both the specification and the executable plan.

Concretely, a stack repository is expected to contain:

1. One manifest (JSON) listing every module it composes, each as a reference to a real, independently publishable repository — never a vendored code copy. See docs/manifest-schema.md for the exact schema (ellmos-stack-manifest-v1) shared across all stacks in this catalog.
2. A thin installer that reads the manifest and clones/wires the listed modules — no stack-specific logic beyond "read manifest, act on it."
3. A README that explains the stack's purpose, the role of each module, and how the modules' declared provides/consumes capabilities connect to each other (a small wiring diagram is usually the clearest way to show this).

This keeps every stack's own repository small: it is composition and documentation, not a monorepo of copied source.

Adding a new stack to this catalog

1. Publish the stack's own repository under ellmos-ai with a manifest that follows ellmos-stack-manifest-v1 (see docs/manifest-schema.md).
2. Add one row to the Catalog table above.
3. If the ellmos-ai .github organization profile lists stacks, link the new stack there too.

License

MIT

---

Haftung / Liability

Dieses Projekt ist eine unentgeltliche Open-Source-Schenkung im Sinne der §§ 516 ff. BGB. Die Haftung des Urhebers ist gemäß § 521 BGB auf Vorsatz und grobe Fahrlässigkeit beschränkt.