A structured development methodology for AI coding agents, inspired by Qian Xuesen's engineering cybernetics. Closed-loop feedback control keeps plans aligned with reality. Pure Markdown, any platform.

Loom is an AI harness, that can be used with local or cloud LLMs, for complex tasks. It decomposes work, drives execution through a verification harness, and keeps models on track with structured state instead of history. It can route between thinking and acting models, verifies outputs, and exposes an APP/TUI/API/CLI/MCP for both humans & agents.

Agentic AI refers to AI systems capable of autonomous decision-making, planning, and executing tasks based on goals—acting like intelligent agents. These systems combine LLMs with tools, memory, and feedback loops to complete complex workflows with minimal human input.

Turn requirements into executable Linear issues — with explicit dependencies, design boundaries, and interactive validation.

Instantly decompose Azure DevOps work items into hierarchical trees and streamline agile planning in a single action! 🚀

Krako 2.0 – Energy-efficient, triadic multi-tier inference infrastructure enabling adaptive routing across heterogeneous edge–cloud nodes.

Make any LLM 10x smarter. Recursive DAG decomposition + temporal knowledge graph = free models that match expensive ones. 30/30 tasks passing at $0.0006 total cost.

AI-Driven Project Decomposition and Implementation

Cursor skills and harness-oriented workflows for project intake, problem framing, and AI-ready task decomposition.

Autonomous coding agent harness built on Claude Code SDK — implements Anthropic's long-running agent pattern: scan → plan → code → validate → git-commit in a loop. Turn requirements into working code with one command.

[Under Review] TDNet (Degradation-aware Comprehensive Task Decomposition for Joint Rain and Haze Removal)

Intelligent travel planning with multi-agent AI orchestration using n8n

🏛️ TipAI Architect — 三方AI答辩式方案架构师。用户提需求→裁判调研设计→正方vs反方辩论→裁决→拆解为原子开发任务

🧠 Self-organizing AI agent with dendrite spawning - decomposes complex goals into neurons, spawns parallel sub-tasks, and aggregates results. Features smart parameter extraction, context isolation, and automatic data compaction. Built for Strava analytics with extensible tool system.

Automatically decompose complex prompts into orchestrated, executable workflows for Claude Code. Research-backed prompt chaining and task decomposition.

Claude 规划技能：把一句话目标转成可直接粘贴给 Codex 的有界任务单（Goal / Scope / Acceptance Criteria / Validation 九段式），让规划模型负责切片，让 Codex 专注执行。

AI-powered task organizer with intelligent decomposition, skill gap analysis, and JIT learning

Enable deterministic distributed execution for AI workflows with replay-safe billing, contention safety, and trust-driven scheduling.