Reading
A gated workflow for AI coding agents targeting production-level development, which structurally manages the complete development process from requirements to delivery through the core mechanism of "no implementation without approved plans".
Section 01
PlanGate is a gated workflow framework for AI coding agents targeting production-level development. Its core mechanism is "no implementation without approved plans", aiming to structurally manage the complete development process from requirements to delivery, and solve key issues in AI coding such as disconnect between planning and implementation, ambiguous approval boundaries, and post-hoc verification.
Section 02
The popularity of AI coding tools has improved development efficiency, but also exposed new risks: AI implementations often outpace plans, approval and responsibility boundaries are ambiguous, verification becomes a post-hoc remedy, context and evaluation standards rely excessively on conversations or personal memory, and completion criteria are scattered across PRs and dialogue records. PlanGate was born to address this, placing AI within a development constraint harness with clear approval boundaries, explicit deliverables, and standard verification processes.
Section 03
PlanGate's core design philosophy is "AI cannot write a single line of code without an approved plan". Its differences from traditional Harness engineering include:
Section 04
PlanGate uses a multi-level gated mechanism:
Section 05
docs/working/TASK-XXXX/Transfer state to the file system; the task directory includes plan.md (plan), review.md (review records), verification.md (verification results), and handoff.md (handoff documents), bringing benefits such as auditability, recoverability, collaboration, and knowledge accumulation.
Section 06
Introduces three-layer separation of Workflow, Skill, and Agent:
Supports the combined use of Claude Code and Codex CLI:
| Tool | Entry File | Dedicated Configuration |
|---|---|---|
| Claude Code | CLAUDE.md | .claude/ |
| Codex CLI | AGENTS.md | .codex/ |
| Shared | docs/ai/project-rules.md | docs/, scripts/ |
Section 07
Progressive strategy:
PlanGate is an attempt to evolve AI-assisted development from "personal tools" to "team collaboration frameworks". It solves core issues through gated mechanisms, state persistence, and multi-level verification, allowing teams to enjoy AI efficiency while maintaining control and confidence in the development process.
Keep going with more reads from the same topic.
Splinter is a minimalist, high-performance key-value (KV) and vector storage system enabling zero-latency inter-process communication via shared memory and atomic operations. With only 766 lines of core code, it supports millions of operations per second and 768-dimensional vector storage, offering a new architectural approach for local LLM inference and data-intensive applications.
Folkering OS is the world's first AI-native bare-metal operating system, entirely written in Rust no_std without relying on Linux, POSIX, or libc. It can generate commands from scratch, compile them into WASM, and run them in 10 seconds, achieving true self-evolution.
This project explores how to use Large Language Models (LLMs) to detect logical vulnerabilities in Ethereum smart contracts, especially business logic flaws that are difficult for traditional static analysis tools to capture, such as economic manipulation and execution order errors.
This article introduces an open-source project for building modern large language models from scratch. The project evolves from a GPT-2-style basic architecture to a Llama 2/3-style production-level implementation, covering complete tutorial implementations of key technologies such as RMSNorm, RoPE, SwiGLU, GQA, and MoE.