v0.4.0 — Agent Failure Risk, Audit Evidence, and Explainable Fix PRs

RepoReady v0.4.0 — Agent Failure Risk, Audit Evidence, and Explainable Fix PRs

RepoReady is evolving from an AI coding agent readiness scanner into an AI Coding Agent Readiness Audit + Fix Platform.

This release focuses on making reports more trustworthy, fixes more reviewable, and the project more useful for repositories that want to work better with Codex, Claude Code, Cursor, and human contributors.

Install

npx @shidesheng0218/repo-ready@latest
Generate a Chinese report:
npx @shidesheng0218/repo-ready@latest --lang zh
Generate Markdown or JSON:
npx @shidesheng0218/repo-ready@latest --markdown
npx @shidesheng0218/repo-ready@latest --json
Preview fix plan:
npx @shidesheng0218/repo-ready@latest fix --plan
Preview generated fixes:
npx @shidesheng0218/repo-ready@latest fix --dry-run
What is RepoReady?
RepoReady helps your repository become more ready for:
Codex
Claude Code
Cursor
AI coding agents
Open-source contributors
It checks whether a repository has the instructions, validation commands, documentation, context quality, and safety boundaries needed for AI-assisted development.
Highlights
Agent Failure Risk
RepoReady now predicts where AI coding agents are most likely to fail inside a repository.
Risk categories include:
Context Confusion
Validation Gap
Safety Boundary Risk
Onboarding Gap
Scope Drift Risk
Each risk includes:
risk level
score
why agents may fail
evidence
mitigation
fixability
Audit-grade Evidence Chain
Reports now include stronger evidence fields:
status
severity
category
source
impact
fixability
suggested fix
This makes RepoReady reports more explainable and easier to trust.
Strategy Layer 2.0
RepoReady now provides a clearer strategic summary:
posture
risk level
readiness gap
evidence confidence
priority actions
automation boundary
growth plan
recommended path
The recommended path is split into:
Now
Next
Later
Explainable Fix PRs
Fix PRs are now more professional and easier to review.
Generated PR bodies include:
Why this PR exists
What RepoReady detected
Agent Failure Risks
What this PR changes
What needs human review
Suggested validation commands
Each generated change can include a reason and risk level.
Better Web Report Storytelling
The web report page now follows a clearer narrative:
Current State
Why Agents May Fail
Fix Path
Share / Growth
The goal is to make reports more useful, more shareable, and easier to understand.
Agent Ready Index
This release adds a generated Agent Ready Index article:
docs/agent-ready-index-2026-06.md
It includes:
executive summary
key findings
top repositories
common blockers
failure risk distribution
methodology
Chinese summary
launch copy for Hacker News, Reddit, X, V2EX, 掘金, and 知乎
Encoding Hardening
Fixed Chinese encoding issues in Markdown output and index article generation.
Added regression tests to prevent Chinese labels from becoming corrupted.
Commands
npx @shidesheng0218/repo-ready@latest
npx @shidesheng0218/repo-ready@latest --lang zh
npx @shidesheng0218/repo-ready@latest --markdown
npx @shidesheng0218/repo-ready@latest --json
npx @shidesheng0218/repo-ready@latest fix --plan
npx @shidesheng0218/repo-ready@latest fix --dry-run
Safety
RepoReady remains safe by default:
It does not execute repository scripts.
It does not upload private local source by default.
It only writes files when explicitly requested.
Fix PRs are reviewable and are not merged automatically.
Dangerous scripts are flagged for manual review.
Validation
This release was verified with:
npm run lint
node --test tests\core.test.js tests\advanced-report.test.js
npm run web:build
npm pack --dry-run --cache .\.npm-cache
Links
GitHub: https://github.com/shidesheng0218/repo-ready
npm: https://www.npmjs.com/package/@shidesheng0218/repo-ready
中文简介
RepoReady v0.4.0 将项目从“AI 编程代理适配扫描器”进一步升级为 AI 编程代理适配审计与修复平台。
这一版重点增强：
Agent 失败风险预测
审计级证据链
策略层 2.0
可解释 Fix PR
Web 报告页叙事
Agent Ready Index 传播报告
中文编码稳定性
RepoReady 的目标是帮助 GitHub 仓库更适合 Codex、Claude Code、Cursor 等 AI 编程代理协作，同时让项目更容易被贡献者理解和参与。