TypeScript AI agent orchestration framework with dynamic workflows. Describe the goal, not the graph: a coordinator plans the task DAG at runtime and runs it on any LLM (Claude, ChatGPT, Gemini, DeepSeek, or local models).

AI-native agent harness for coding workflows by python: multi-model LLM orchestration, stateful sessions, tool governance, traceable delivery, and provider routing for GPT, Claude, DeepSeek, Qwen, Kimi, GLM, and MiniMax.

Run Claude Code dynamic workflows on a local Codex (GPT) backend, plus an interactive run viewer

Codex-first Ultracode skill for dynamic coding workflows

Claude Code–style dynamic workflows for Pi: code-mode subagents with real model routing, journaled resume, git-worktree isolation, cost accounting, an interactive /workflows TUI, an /ultracode standing opt-in, and deep research.

Dynamic Workflows for hermes

Open Dynamic Workflows: TypeScript CLI runtime for Claude Code-style dynamic workflows across Codex, Claude Code, Gemini, Qwen, Kimi, and custom coding-agent CLIs.

An abstraction library for building domain-specific intelligent agents based on Large Language Models (LLMs). LLMAgent provides a core architecture and behavior definitions that simplify the development of specialized agents by leveraging the AgentForge framework for dynamic workflow orchestration.

Provide a comprehensive collection of system design concepts and examples using `Mermaid diagrams` and `draw.io`.

A community-curated collection of real-world Claude Code Dynamic Workflows — multi-agent orchestration scripts that install as slash commands, shared by the people running them in production.

Local durable harness for AI-first workflows

AgentForge is a powerful and flexible signal-driven workflow framework designed for building intelligent, dynamic, and adaptive systems.

Open-source dynamic multi-agent workflows for AI coding agents - the script-as-orchestrator engine behind Claude Code dynamic workflows & ultracode, for OpenCode, Codex, Antigravity & VS Code. Local-first, bring your own model (Anthropic/OpenAI/Ollama), MIT.

Evolving agent harnesses: a research program on how far N orchestrated calls of a small model can rival a frontier model. We evolve the harness (structure + prompts) with reflective optimizers + a verified-acceptance gate.

Visualize Claude Code dynamic workflows from your terminal — a fan-out→barrier→reduce DAG, a gantt timeline, live concurrency, and a self-contained interactive HTML report. Zero runtime deps.

Open-Source implementation of Claude dynamic workflows in opencode.

Dynamic workflow tooling for Codex: planner/replanner orchestration, Codex CLI process swarms, broker evidence, and benchmark reports.

A multi-agent equity-analysis Claude Code Dynamic Workflow — analysts → bull/bear debate → trader → risk → PM, with deterministic Markdown + HTML export. Research only, not financial advice.

Hermes Agent plugin prototype for Claude Code-style dynamic workflows

Dynamic Hierarchical Workflows via Generative Nextflow