Atlas / Skills / xu-xiang / Skill Stocktake

Skill StocktakeSAFE

skills/xu-xiang/skill-stocktake

everything-claude-code 中文翻译项目:完整的 Claude Code 配置集合(agents, skills, hooks, commands, rules, MCPs)。源自 Anthropic 黑客松获胜者的实战配置,助力中文工程师高效理解与使用 Claude Code。

Verdict
SAFE
Grade
B
Trust score
89 /100
Version
—
Hosts
—
License
MIT
Stars
1,988
01

Overview

everything-claude-code 中文翻译项目:完整的 Claude Code 配置集合(agents, skills, hooks, commands, rules, MCPs)。源自 Anthropic 黑客松获胜者的实战配置,助力中文工程师高效理解与使用 Claude Code。

Read from source at commit 012f775295c3OBSERVED · 2026-10-12
02

What it tells the agent

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
description: "用于审计Claude技能和命令的质量。支持快速扫描(仅变更技能)和全面盘点模式,采用顺序子代理批量评估。"
origin: ECC
---

# skill-stocktake

斜杠命令 (`/skill-stocktake`),用于使用质量检查清单 + AI 整体判断来审核所有 Claude 技能和命令。支持两种模式:用于最近更改技能的快速扫描,以及用于完整审查的全面盘点。

## 范围

该命令针对以下**相对于调用命令所在目录**的路径:

| 路径 | 描述 |
|------|-------------|
| `~/.claude/skills/` | 全局技能(所有项目) |
| `{cwd}/.claude/skills/` | 项目级技能(如果目录存在) |

**在第 1 阶段开始时,该命令会明确列出找到并扫描了哪些路径。**

### 针对特定项目

要包含项目级技能,请从该项目根目录运行:

```bash
cd ~/path/to/my-project
/skill-stocktake
```

如果项目没有 `.claude/skills/` 目录,则只评估全局技能和命令。

## 模式

| 模式 | 触发条件 | 持续时间 |
|------|---------|---------|
| 快速扫描 | `results.json` 存在(默认) | 5–10 分钟 |
| 全面盘点 | `results.json` 不存在,或 `/skill-stocktake full` | 20–30 分钟 |

**结果缓存:** `~/.claude/skills/skill-stocktake/results.json`

## 快速扫描流程

仅重新评估自上次运行以来发生更改的技能(5–10 分钟)。

1. 读取 `~/.claude/skills/skill-stocktake/results.json`
2. 运行:`bash ~/.claude/skills/skill-stocktake/scripts/quick-diff.sh \   ~/.claude/skills/skill-stocktake/results.json`
   (项目目录从 `$PWD/.claude/skills` 自动检测;仅在需要时显式传递)
3. 如果输出是 `[]`:报告“自上次运行以来无更改。”并停止
4. 使用相同的第 2 阶段标准仅重新评估那些已更改的文件
5. 沿用先前结果中未更改的技能
6. 仅输出差异
7. 运行:`bash ~/.claude/skills/skill-stocktake/scripts/save-results.sh \   ~/.claude/skills/skill-stocktake/results.json <<< "$EVAL_RESULTS"`

## 全面盘点流程

### 第 1 阶段 — 清单

运行:`bash ~/.claude/skills/skill-stocktake/scripts/scan.sh`

脚本枚举技能文件,提取 frontmatter,并收集 UTC 修改时间。
项目目录从 `$PWD/.claude/skills` 自动检测;仅在需要时显式传递。
从脚本输出中呈现扫描摘要和清单表:

```
Scanning:
  ✓ ~/.claude/skills/         (17 files)
  ✗ {cwd}/.claude/skills/    (not found — global skills only)
```

| 技能 | 7天使用 | 30天使用 | 描述 |
|-------|--------|---------|-------------|

### 第 2 阶段 — 质量评估

启动一个 Task 工具子代理(**Explore 代理,模型:opus**),提供完整的清单和检查清单。
子代理读取每个技能,应用检查清单,并返回每个技能的 JSON:

`{ "verdict": "Keep"|"Improve"|"Update"|"Retire"|"Merge into [X]", "reason": "..." }`

**分块指导:** 每个子代理调用处理约 20 个技能,以保持上下文可管理。在每个块之后将中间结果保存到 `results.json` (`status: "in_progress"`)。

所有技能评估完成后:设置 `status: "completed"`,进入第 3 阶段。

**恢复检测:** 如果在启动时找到 `status: "in_progress"`,则从第一个未评估的技能处恢复。

每个技能都根据此检查清单进行评估:

```
- [ ] Content overlap with other skills checked
- [ ] Overlap with MEMORY.md / CLAUDE.md checked
- [ ] Freshness of technical references verified (use WebSearch if tool names / CLI flags / APIs are present)
- [ ] Usage frequency considered
```

判定标准:

| 判定 | 含义 |
|---------|---------|
| Keep | 有用且最新 |
| Improve | 值得保留,但需要特定改进 |
| Update | 引用的技术已过时(通过 WebSearch 验证) |
| Retire | 质量低、陈旧或成本不对称 |
| Merge into \[X] | 与另一技能有大量重叠;命名合并目标 |

评估是**整体 AI 判断** — 不是数字评分标准。指导维度:

* **可操作性**:代码示例、命令或步骤,让你可以立即行动
* **范围契合度**:名称、触发器和内容保持一致;不过于宽泛或狭窄
* **独特性**:价值不能被 MEMORY.md / CLAUDE.md / 其他技能取代
* **时效性**:技术引用在当前环境中有效

**原因质量要求** — `reason` 字段必须是自包含且能支持决策的:

* 不要只写“未更改” — 始终重述核心证据
* 对于 **Retire**:说明 (1) 发现了什么具体缺陷,(2) 有什么替代方案覆盖了相同需求
  * 差:`"Superseded"`
  * 好:`"disable-model-invocation: true already set; superseded by continuous-learning-v2 which covers all the same patterns plus confidence scoring. No unique content remains."`
* 对于 **Merge**:命名目标并描述要集成什么内容
  * 差:`"Overlaps with X"`
  * 好:`"42-line thin content; Step 4 of chatlog-to-article already covers the same workflow. Integrate the 'article angle' tip as a note in that skill."`
* 对于 **Improve**:描述所需的具体更改(哪个部分,什么操作,如果相关则说明目标大小)
  * 差:`"Too long"`
  * 好:`"276 lines; Section 'Framework Comparison' (L80–140) duplicates ai-era-architecture-principles; delete it to reach ~150 lines."`
* 对于 **Keep**(快速扫描中仅 mtime 更改):重述原始判定理由,不要写“未更改”
  * 差:`"Unchanged"`
  * 好:`"mtime updated but content unchanged. Unique Python reference explicitly imported by rules/python/; no overlap found."`

### 第 3 阶段 — 摘要表

| 技能 | 7天使用 | 判定 | 原因 |
|-------|--------|---------|--------|

### 第 4 阶段 — 整合

1. **Retire / Merge**:在用户确认之前,按文件呈现详细理由:
   * 发现了什么具体问题(重叠、陈旧、引用损坏等)
   * 什么替代方案覆盖了相同功能(对于 Retire:哪个现有技能/规则;对于 Merge:目标文件以及要集成什么内容)
   * 移除的影响(是否有依赖技能、MEMORY.md 引用或受影响的工作流)
2. **Improve**:呈现具体的改进建议及理由:
   * 更改什么以及为什么(例如,“将 430 行压缩至 200 行,因为 X/Y 部分与 python-patterns 重复”)
   * 用户决定是否采取行动
3. **Update**:呈现已检查来源的更新后内容
4. 检查 MEMORY.md 行数;如果超过 100 行,则建议压缩

## 结果文件模式

`~/.claude/skills/skill-stocktake/results.json`:

**`evaluated_at`**:必须设置为评估完成时的实际 UTC 时间。
通过 Bash 获取:`date -u +%Y-%m-%dT%H:%M:%SZ`。切勿使用仅日期的近似值,如 `T00:00:00Z`。

```json
{
  "evaluated_at": "2026-02-21T10:00:00Z",
  "mode": "full",
  "batch_progress": {
    "total": 80,
    "evaluated": 80,
    "status": "completed"
  },
  "skills": {
    "skill-name": {
      "path": "~/.claude/skills/skill-name/SKILL.md",
      "verdict": "Keep",
      "reason": "Concrete, actionable, unique value for X workflow",
      "mtime": "2026-01-15T08:30:00Z"
    }
  }
}
```

## 注意事项

* 评估是盲目的:无论来源如何(ECC、自创、自动提取),所有技能都应用相同的检查清单
* 归档 / 删除操作始终需要明确的用户确认
* 不按技能来源进行判定分支
03

Trust audit

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

LayerWhat it checksResult
L0Provenance & inventoryPASS
L1Static analysis of the codeNA
L2Instruction surface (what it tells the agent)PASS
L3Class-specific surfacePASS
L4Behavioural (sandbox)SKIPPED

What the source does

Filesystem
none-observed
Network
none-observed
Shell
none-observed
Dependencies
pinned
Secrets in source
none-found

Findings (1)

LOWInventory / provenance · skill.no_frontmatter · CWE-1104
SKILL.md:1
Why it matters. SKILL.md lacks name/description frontmatter

Gates applied: no_behavioural_pass.

Audited 2026-10-12 · audit v0.4.1 · source sha 012f775295c3full audit observations/trust-audit/skill/xu-xiang__skill-stocktake.json · Report an issue / request a re-scan
04

Audit history

Every audit this skill has had.

DateSourceVerdictGradeScoreChange
2026-10-12012f775295c3SAFEB89first audit
05

Questions

What does the Skill Stocktake skill do?

everything-claude-code 中文翻译项目:完整的 Claude Code 配置集合(agents, skills, hooks, commands, rules, MCPs)。源自 Anthropic 黑客松获胜者的实战配置,助力中文工程师高效理解与使用 Claude Code。

Is Skill Stocktake safe to install?

The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean skill reads B.

What can Skill Stocktake access on my machine?

The audit observed no filesystem, network or shell use at all in its source.

How current is this page?

The grade is for one exact copy of the source (012f775295c3), read on 2026-10-12. The repository is watched, and a new audit runs when it changes — this is the first audit.

Advertisement