claude-code-feature-validation SKILL.md skillA

claude-code-feature-validation from esonhugh/claude-code is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.

Install it, pinned
mdr add esonhugh/claude-code/claude-code-feature-validation@v0.1.0

Pins v0.1.0 in mdr.lock. Pin the fingerprint instead.

Audit grade
A, 17 of 17 checks passed
Kind
SKILL.md skill
Repository
esonhugh/claude-code
Path
.claude/skills/claude-code-feature-validation/SKILL.md
Versions recorded
2
Latest version
v0.1.0, committed Jul 23, 2026
First version recorded
Jul 18, 2026
Size
13,895 bytes
Repository stars
43
License
none declared

Use this skill automatically whenever validating a Claude Code repository change: a new feature, modified feature, bug fix, refactor, CLI command, tool, Agent/Workflow behavior, task lifecycle, React Ink/TUI interaction, packaging, or official/local compatibility. Runtime changes, including internal or not directly user-visible changes, require both related bun tests and current-binary scripted tmux coverage of affected entry points, adjacent consumers, and side effects. Also trigger for…

Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.

What the file says

# Claude Code Feature Validation

为 Claude Code 新增或修改 feature 后,根据真实改动和风险选择最小但充分的验证路径,并输出可审计结论。不要把“命令退出 0”“测试通过”或“runtime done”单独等同于 feature 已通过验收。

## 核心原则

1. **先理解改动,再选择验证。** 阅读相关 diff、实现、调用链、测试和项目脚本,不对未读代码设计具体断言。
2. **从 feature 契约和调用链反推 assertions。** 同时验证内部契约、真实 runtime 入口、相邻消费者、用户可见行为和副作用,而不只是实现细节。
3. **验证层级与风险匹配。** 默认从最小范围开始;只有当前层不能证明行为时才升级。
4. **测试和运行时证据互补。** 自动化测试证明可重复逻辑;真实 CLI/TUI 验收证明进程内集成和用户可见行为。
5. **所有 runtime feature 必须双重通过。** 无论改动是否直接用户可见,只要进入 `built-claude` 并影响功能或共享 runtime 路径,就必须同时通过相关 `bun test` 与本轮 binary 的 scripted tmux 交互验收。内部改动要通过其真实调用入口触发,并覆盖直接受影响的相邻组件和副作用。任一失败、仍在运行或证据不足时,整体 feature 必须标记为 `failed`、`running` 或 `not covered`;build、`make release-check`、L4 或 parent-side 工具均不能替代交互证据。
6. **结论必须绑定证据。** 缺少必要证据时标记 `not covered`,仍运行标记 `running`,不得用推测补齐。
7. **不制造额外副作用。** 不 commit、push、创建 PR、发布、修改共享配置或删除用户状态,除非用户明确授权。
8. **尊重当前工作区。** 不覆盖或清理既有未提交改动。验证前后检查 Git 可见状态,将构建产物与被测运行副作用分开。

## 第一步:建立变更契约

开始验证前,写出精简 validation scope:

- changed behavior:新增、修改或修复了什么;
- affected surfaces:纯函数、状态、组件、命令、tool schema、CLI 入口、TUI、Agent/Workflow、构建或发布产物;
- user-visible contract:用户输入、状态迁移、输出、错误行为和副作用;
- regressions at risk:最可能被破坏的相邻行为;
…

Read the whole file at its exact version.

How to install

The exact contents, by fingerprint
mdr add esonhugh/claude-code/claude-code-feature-validation@sha256:305c8203eacc6490

The line at the top of this page pins the label v0.1.0, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.

Badge

mdr badge

[![mdr](https://markdownregistry.com/badge/art_kulpfnak6vu5hn7d.svg)](https://markdownregistry.com/a/art_kulpfnak6vu5hn7d)

1 badge image fetches in 30 days. What the badge shows.

Versions

The registry has recorded 2 versions of this file, between Jul 18, 2026 and Jul 23, 2026, about one every 4 days.

versioncommittedcommitsizeaudit
v0.1.0 latest2026-07-23 aac126c 13,895 BA view · diff
v0.1.02026-07-18 375159f 11,695 BA view

Audit of the latest version

A  17 of 17 checks passed. Fixed automatic checks with no AI involved, so the same checks give the same answer every run.
  • pass: Frontmatter block present
  • pass: Frontmatter declares a name
  • pass: Frontmatter declares a description
  • pass: Size between 200 bytes and 200 KB (13895 bytes)
  • pass: No zero-width or bidi control characters
  • pass: No instruction hidden inside an HTML comment
  • pass: No link to an exfiltration or paste host
  • pass: No credential-shaped string
  • pass: No instruction to send local credentials anywhere
  • pass: No text hidden with inline styles
  • pass: No prompt-injection phrasing
  • pass: No curl or wget piped into a shell
  • pass: No recursive delete of root, home or parent
  • pass: No instruction to read or print local credentials
  • pass: No base64 blob over 200 characters
  • pass: No link to a raw IP address
  • pass: No script tag

Source

GitHub

esonhugh/claude-code · 43 stars · license none declared · pushed 2026-10-20 · branch master

API

GET https://markdownregistry.com/api/v1/artifacts/art_kulpfnak6vu5hn7d
GET https://markdownregistry.com/api/v1/resolve?ref=esonhugh/claude-code/claude-code-feature-validation
GET https://markdownregistry.com/api/v1/blob/305c8203eacc6490d6590415403be6713b35597df8e8aad34ba50f197dda84a0

Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.

More from esonhugh/claude-code

claude-agent-workflow-validation SKILL.md skill
esonhugh/claude-code · .claude/skills/claude-agent-workflow-validation/SKILL.md · This skill should be used automatically when the user asks for strict binary-side validation of Agent, Workflow…
v0.3.0 · audit A · 43 stars
claude-network-debug SKILL.md skill
esonhugh/claude-code · .claude/skills/claude-network-debug/SKILL.md · This skill should be used when the user asks about Claude Code “代理不生效”, “OAuth 网络失败”, “抓流”, “CONNECT”, “SSE/WebSocket”…
v0.1.0 · audit A · 43 stars
claude-runtime-debug SKILL.md skill
esonhugh/claude-code · .claude/skills/claude-runtime-debug/SKILL.md · This skill should be used when the user asks to diagnose Claude Code CLI/runtime behavior, including “运行时调试”…
v0.2.0 · audit A · 43 stars
claude-source-binary-analysis SKILL.md skill
esonhugh/claude-code · .claude/skills/claude-source-binary-analysis/SKILL.md · This skill should be used when the user asks to statically analyze authorized Claude Code source or…
v0.2.0 · audit A · 43 stars
release-validation SKILL.md skill
esonhugh/claude-code · .claude/skills/release-validation/SKILL.md · Validates release readiness with related bun tests, make release-check, current-binary scripted tmux interactions, and…
v0.1.0 · audit A · 43 stars
CLAUDE.md CLAUDE.md file
esonhugh/claude-code · CLAUDE.md
git:20260723.44dceca · audit A · 43 stars
claude-api SKILL.md skill
esonhugh/claude-code · src/skills/bundled/claude-api/SKILL.md
git:20260331.4178dd7 · audit C · 43 stars
verify SKILL.md skill
esonhugh/claude-code · src/skills/bundled/verify/SKILL.md · Verify a code change by running the app and checking the user-facing outcome.
git:20260331.4178dd7 · audit B · 43 stars

Every file in esonhugh/claude-code

Browse by kind, by grade A, or by owner.