Skip to content

v0.2.0 – Agent Ready Spec, Policy Layer, and Fix Plans

Choose a tag to compare

@shidesheng0218 shidesheng0218 released this 29 Jun 06:25

你在 GitHub Release 的 Write 里面直接复制下面这一整段即可:

# RepoReady v0.2.0

RepoReady v0.2.0 upgrades the project from an AI-agent repository scanner into an early **AI Agent Readiness infrastructure layer**.

RepoReady helps your repository become easier for Codex, Claude Code, Cursor, and human contributors to understand, validate, and safely modify.

## Install

```bash
npx @shidesheng0218/repo-ready@latest

Chinese report:

npx @shidesheng0218/repo-ready@latest --lang zh

What's new

Agent Ready Spec

RepoReady now includes an Agent Ready Repository Standard.

npx @shidesheng0218/repo-ready@latest spec
npx @shidesheng0218/repo-ready@latest spec --lang zh

This defines what an AI-agent-ready repository should include:

  • agent instructions
  • install / dev / test / build / lint commands
  • README onboarding
  • safety boundaries
  • clean repository context
  • reviewable agent tasks

Policy Layer

RepoReady now supports policy initialization and policy checks.

Generate a default policy:

npx @shidesheng0218/repo-ready@latest policy init

Write policy config:

npx @shidesheng0218/repo-ready@latest policy init --write

Check policy compliance:

npx @shidesheng0218/repo-ready@latest policy check

This is the first step toward making RepoReady useful for teams that want enforceable readiness rules before AI coding agents modify repositories.

Fix Plan

RepoReady can now classify generated fixes into:

  • safe automatic fixes
  • fixes that need review
  • manual-only fixes

Run:

npx @shidesheng0218/repo-ready@latest fix --plan

Apply only safe automatic fixes:

npx @shidesheng0218/repo-ready@latest fix --apply-safe

This makes RepoReady safer and more practical for real repositories.

Existing features

RepoReady also includes:

  • AI Agent Ready score
  • Contributor Ready score
  • Context Quality score
  • Safety score
  • Code Quality score
  • evidence chain
  • score breakdown
  • bilingual reports
  • doctor mode
  • agent task generation
  • context pack generation
  • GitHub Action wrapper
  • fix dry-run / write / branch / PR prototype

Example commands

npx @shidesheng0218/repo-ready@latest
npx @shidesheng0218/repo-ready@latest --lang zh
npx @shidesheng0218/repo-ready@latest doctor
npx @shidesheng0218/repo-ready@latest tasks
npx @shidesheng0218/repo-ready@latest context --dry-run
npx @shidesheng0218/repo-ready@latest fix --dry-run
npx @shidesheng0218/repo-ready@latest fix --plan
npx @shidesheng0218/repo-ready@latest policy check

Why this release matters

AI coding agents work better when repositories provide clear instructions, validation commands, clean context, and safety boundaries.

RepoReady v0.2.0 moves toward becoming the preflight layer for AI coding agents:

Before an agent changes code, RepoReady checks whether the repository is understandable, testable, safe, and actionable.

中文一句话:

RepoReady 帮你把仓库整理成 AI 编程代理和真人贡献者都更容易理解、更容易验证、更安全协作的状态。

Links

GitHub:

https://github.com/shidesheng0218/repo-ready

npm:

https://www.npmjs.com/package/@shidesheng0218/repo-ready