agents-md-template

This repo hands your AI coding agent a rulebook and your CI the checks that enforce it, catching the shortcuts agents take by default — skipped tests, widened permissions, fail-open model output.

Build an internal tool (with AI) as one self-contained HTML file — no server, no install, no network. A template for AI coding agents: describe the tool, swap one file, ship the result.

A keel for human + AI project management. Markdown as the database, Git as the audit log, Python as the constitution enforcer — decisions never forgotten, AI never overreaches, drafts expire. | 给「人 + AI 协作的项目管理」装一根龙骨。用 Markdown 当数据库，用 Git 当审计日志，用 Python 脚本当宪法执行者 —— 决策不失忆，AI 不越权，草稿会过期。

The operating system for AI-built software: one rulebook, a skill library, project memory on disk, and gates that hold. Copy it before you start.

Model-agnostic AGENTS.md templates and project instructions for AI coding agents such as Codex, Claude Code and Gemini CLI.

"code this idea" — turns a plan or idea into a complete, AI-coding-agent-ready docs set (AGENTS.md + linked architecture/decisions/roadmap/product/design docs). Interviews you with grounded recommendations instead of a fixed template. Works with Claude Code, Codex, Cursor, and other AGENTS.md-compatible agents.

Stay in control of AI-written code without drowning in implementation details. Works with Claude Code, Codex, and any AGENTS.md-compatible agent.

A reference operating framework for coding agents.

Single-file operating standard for Claude Code: verification discipline, honest pushback, calibrated autonomy

Compile ordered Markdown policy modules into one deterministic, verifiable AGENTS.md. The target can be global, repository-local, nested, or another explicit path. The runtime has no dependencies, performs no network access, invokes no shell, and uses no LLM.

Make Codex smarter with persistent project memory, AGENTS.md templates, smart routing, and session continuity for AI coding agents.

A repository-native governance framework for defining how humans and AI agents work together across projects, tools, and workflows.

One VS Code sidebar for the memory, rules and skills of every AI coding agent — Claude Code, Cursor, Kiro, Antigravity and AGENTS.md. Shared prompt/skill libraries for your team, one-click skill install, and safe recovery when a project folder is renamed.

30 runs testing whether AGENTS.md changes coding-agent behaviour, and an honest post-mortem of why the design couldn't answer it.

Wire-format specifications for signing AI instruction files — SKILL.md bundles, agent definitions, CLAUDE.md/AGENTS.md - and the scripts that ship with them.

AI-assisted production workspace: agent-navigable knowledge base, versioned runbooks, pro git discipline. Claude Code native, AGENTS.md compatible.

Turn the rules in your CLAUDE.md into deterministic checks, and see whether a real Claude Code run respects them

AI-Zoll prepares, structures, and maintains the context that AI coding agents need to build software effectively. It does not write your application code — Claude Code, Cursor, Codex, and Copilot remain the builders. Zoll prepares the environment in which they build.

A Codex agent workflow that extracts songs from screenshots and finds verified YouTube matches.

Context-engineering kit for AI coding agents.