AI Meta
Repository: github.com/noetl/ai-meta
ai-meta is the coordination layer for the NoETL multi-repo ecosystem. It provides three things:
- Deterministic ecosystem state — all NoETL repos pinned as Git submodules at exact SHAs
- Git-tracked shared memory — a persistent knowledge store that survives across sessions and is shared across engineers and AI agents
- Agent infrastructure — shared rules, skills, and profiles that any AI agent (Claude Code, Codex, Copilot, Cursor) can use
Repository Layout
ai-meta/
├── AGENTS.md # Universal rules (all agents, all humans)
├── CLAUDE.md # Claude Code auto-bootstrap entry point
├── .cursorrules # Cursor entry point
├── .github/copilot-instructions.md # GitHub Copilot entry point
│
├── agents/ # SHARED agent infrastructure
│ ├── rules/ # 7 modular rule files
│ ├── skills/ # 4 workflow definitions (memory-add, compact, sync-note, bump-pointer)
│ └── profiles/ # per-agent behavioral profiles (claude, codex)
│
├── .claude/ # Claude Code specific
│ ├── settings.json # permissions, hooks, env
│ ├── rules -> ../agents/rules # symlink to shared rules
│ ├── skills -> ../agents/skills # symlink to shared skills
│ └── agents/ # Claude subagent defs (frontmatter + @import)
│
├── memory/ # Git-tracked shared memory
│ ├── current.md # active working state (session bootstrap)
│ ├── timeline.md # chronological index
│ ├── inbox/YYYY/MM/ # raw timestamped entries
│ ├── compactions/ # periodic consolidated summaries
│ └── archive/YYYY/MM/ # processed entries
│
├── scripts/ # Automation
│ ├── memory_add.sh # create memory entry
│ └── memory_compact.sh # compact inbox -> summary
│
├── playbooks/ # Operational runbooks
│ ├── how_to_use_ai_meta_day_to_day.md
│ ├── cross_repo_change.md
│ └── memory_compaction.md
│
├── sync/ # Cross-repo coordination
│ ├── TEMPLATE.md # structured sync note template
│ └── issues/ # per-issue tracking docs
│
└── repos/ # Git submodules (12 ecosystem repos)
├── noetl/ # core Python engine
├── server/, worker/, gateway/ # Rust services
├── cli/, tools/ # CLI and shared crates
├── ops/ # deployment automation
├── docs/ # this documentation site
├── gui/ # web UI
└── apt/, homebrew-tap/, noetl.io/
In This Section
| Page | What It Covers |
|---|---|
| Overview | What ai-meta is, the memory system, day-to-day workflow |
| Quick Start for Developers | Clone, setup, first memory entry, first sync note |
| Agent Setup | How each AI agent (Claude, Copilot, Cursor, Codex) connects |
| Shared Memory for Teams | Concurrency patterns, team conventions, what is safe vs. coordinated |
| How Agents Work | Deep dive: agent internals, RAG, context assembly, planning loops |
| Indexes and Timelines | How timeline.md, current.md, and topic indexes support retrieval |
| Agent Orchestration | Using NoETL as an agent registry and orchestrator (agents as playbooks) |