task-dag

Opt-in Codex Skill and plugin for reliable coding: reactive failure recovery, elastic agent teams, task DAGs, one canonical writer, and assured delivery.

Directed Acyclic Graph (DAG) task scheduling engine implementing Critical Path Method (CPM) and slack computation.

Directed Acyclic Graph (DAG) task scheduling engine implementing Critical Path Method (CPM) and slack computation.

Cache-aware orchestration for LLM agents. Fork helpers that share cached prefixes, detect cache breaks, and cut token costs by 38%+.

TypeScript framework that turns one goal into a graph of tasks and runs a team of AI agents to complete it: planning, task dependencies (DAG), tool use and coordination, with only three runtime dependencies. Works with Claude, OpenAI and other LLMs. Modified copy of open-multi-agent/open-multi-agent.

Best OpenCode Fork with Task DAGs & Semantic Retrieval 2026 - Stateful Agent Workbench

Multi-agent coordination for Claude Code. Task DAGs, HMAC auth, browser dashboard. Python stdlib only.

OpenCodex: OpenCode fork adding task DAGs, semantic retrieval, TUI proofs, and a stateful agent workbench runtime.

企业级 Deep Research 多智能体平台：动态任务 DAG、可治理长期记忆、可观察可干预执行与降级检索。

手配 (tehai) — controlled, auditable delegation: request → contract-bound task DAG → cheapest capable model → risk-based review → judge → ledger. Deterministic, stdlib-only, LLM backend behind hard guards.

MCP server that turns Codex into a workforce orchestrator: task-DAG workflows, isolated git worktrees, durable supervision across restarts, budgets, roles, acceptance tracking, metrics and an artifact cache over pluggable runtime backends. Zero-config default: DeepSeek Harness + Kimi K3.