Loop Engineering is the practice of designing recurring systems for AI agents and coding agents. Instead of prompting an agent turn by turn, you build a loop that discovers work, delegates it to one or more agents, verifies the result against tests or other deterministic gates, persists state outside the model, decides what happens next, and runs again on a cadence, an event, or until a verifiable goal is reached. It sits above prompt, context, and harness engineering: those improve a single run, while loop engineering governs repeated agent work over time, including budgets, retries, escalation to humans, and stopping conditions.
Personal memory across agents
Agent OS: the agent gets smarter on its own. We just hold the line: Interview-gated, staged evaluation, budgeted evolution loop. MCP server, 14 runtimes: Claude Code, Codex CLI, Gemini CLI, OpenCode, Copilot, Kiro and more.
A control plane with a durable state kernel for long-horizon agents and teams. Keep work moving and improving across sessions, with less human attention.
The long-horizon computer-use harness. Run AI agents across desktop apps and the CLI for extended periods while preserving task state and making reliable progress on complex workflows. Features fresh-context execution, durable verified state, independent auditing, recoverable progress, and native Claude Code / Codex / OpenClaw integration.
The living ecosystem where AI agents complete tasks through workflow loops, improve through iterative execution, are evaluated by mentor agents or humans in the loop, and turn completed work into reusable work experience and data to improve future agents.
Codebase harness + loop engineer
Self-evolving second brain with 35 AI skills, 10 agents, and people CRM. Closed-loop harness: a V-model verification lifecycle where the worker never grades its own homework. Plus paired anti-slop design skills for marketing and product UI. Works with Claude Code, Cursor, Kiro, Gemini CLI, Codex.
LoongFlow is an expert-grade Agent framework for Loop Engineering. Through a Plan-Execute-Summary loop and structured experiential memory, it enables AI to continuously think, execute, reflect, and evolve across complex software engineering, mathematical, and machine learning tasks.
Cut AI context cost without trusting the compressor. Every reduction is reversible, byte-exact recoverable, and carries an auditable receipt. Local-first, works through proxy, MCP, SDK, or agent wrapper.
Bounded-autonomy plugin harness for agents. Intent to implementation: ready, then go.
Hands-on tutorials for building AI agents from scratch. Learn LLM APIs, prompt engineering, tool calling, and the agent loop through practical examples.
The Loop Engine for Claude Code — engineer the loop, not the prompt. 1 router · 9 agents · 16 skills · 4 workflows. Fail-closed gates, test honesty, anti-anchored review.
Benchmarking models as runtime Controllers for Loop Engineering
agent wiki +engineering skills
Deterministic memory layer for your agents
Turns repeatable, domain-agnostic workflows into graph-driven loops.
Skills for agentic coding CLIs: output you can ship with minimal review, on work you can tell was worth doing.
🔁 Build reliable recurring AI-agent systems: 1025 resources, 22 operational patterns, 22 loop contracts, 8 runtime starters, an interactive atlas, and a structured dataset.
论文解析 Agent 系统（Agent Harness + Loop Engineering），实现结构化抽取与闭环验证优化。
Production-grade software delivery for coding-agent teams on Multica — dynamic planning, deterministic execution, verifiable delivery.