An open-source graph engineering runtime that keeps orchestration in TypeScript and delegates semantic work to replaceable Agent runtimes.

[Up-to-date] A curated list of resources on graph-empowered agents and agent-facilitated graph learning (Graphs Meet Agents).

agent wiki +engineering skills

Build stateful agent workflows with typed outputs, reusable tools, session forks, and ordinary TypeScript.

Turns repeatable, domain-agnostic workflows into multi-step graph-driven loops.

🐙 Graph engineering for long-horizon agents — a curated prompt library wiring specialized clean-context roles (executor · supervisor · scout) into a graph, compiled to each host (Claude Code · grok · Cursor · Codex). Arms: loop-graph · quest.

Desktop app for harness engineering, loop engineering, graph engineering—and whatever comes next in local AI-agent workflows.

🕸️ Engineer the organization, not just the agent. 455 curated resources · 9 design layers · 11 sections · 207 papers & preprints — a field guide, CC0 open dataset, and interactive atlas for graph-structured multi-agent systems: roles, topologies, handoffs, work graphs, state, gates, reliability, observability.

Design grounded graphs of governed improvement loops.

Installable graph engineering for Claude Code, Codex, OpenCode, and Cursor — dependency-graph execution with local caching, quality gates, selective retries, and live reports

Design the structures your agents work through — knowledge graphs for memory, task graphs for orchestration. Playbook + agent skill + runnable stdlib-only pipeline.

Graph Engineering for Agent Skills: a specification and toolchain for dynamically discovering context and building observable, testable, and recoverable agent workflows.

Python toolkit for multi-step AI/agent systems as explicit graphs — define nodes/edges, structural validate (V1–V9), Mermaid visualize, pattern init, and skeleton walk. Vendor-agnostic. Runtime agent execute later.

Central dogma of biology as map for Prompt→Loop→Graph→Evolution trajectory of AI agents. Skill framework, research scaffold, practical patterns.

Progress-aware execution for durable, grounded, token-efficient coding agents.

An incident response copilot built with LangGraph, parallel investigation, human-in-the-loop approval, and long-term memory that improves with each resolved incident.

Bounded static DAG contracts, validators, experiments, and adjacent agent architecture boundaries.

The coding-agent workbench where a deterministic scheduler — not the model — owns control flow. Orchestrates Claude Code & Codex through one auditable, resumable control plane.

Build graph-structured multi-agent systems with this collection of research papers, datasets, and design patterns for programmable AI organizations.

An AI coding agent built for developers who refuse black boxes. Define agents in YAML, let the LLM pick models per task, swap tools via capability routing. Flexibility first, lock-in never.