skill-design-benchmark SKILL.md skillA

skill-design-benchmark from wangjs-jacky/jacky-skills is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.

Install it, pinned
mdr add wangjs-jacky/jacky-skills/skill-design-benchmark@git:20260711.dae29b8

Pins git:20260711.dae29b8 in mdr.lock. Pin the fingerprint instead.

Audit grade
A, 17 of 17 checks passed
Kind
SKILL.md skill
Repository
wangjs-jacky/jacky-skills
Path
plugins/evaluators/skill-design-benchmark/SKILL.md
Versions recorded
1
Latest version
git:20260711.dae29b8, committed Jul 11, 2026
Size
4,974 bytes
Repository stars
15
License
MIT

评估一个或一批 Agent Skills 的设计质量,按 8 个有来源、有证据的维度输出 0–100 分、S/A/B/C/D 等级和优先改进项。重点检查触发边界、完成闭环、决策适应、安全恢复、上下文效率、复杂度适配、一致性和真实行为;不奖励文件数量或流程堆砌。用于‘给 skill 打分’、‘扫描全部 skills’、‘benchmark skill 设计’、‘比较哪些 skills 值得优化’等场景;只评估,不修改被评 Skill。.

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

# Skill Design Benchmark

以最少上下文,对 Agent Skill 的设计质量做可追溯评分。评当前证据,不评作者意图,也不评领域知识多少。

## 不变量

- **默认只读**:检查文件、脚本、测试和现有运行记录;不修改被评 Skill。
- **每个分数必须附证据**:引用目标文件路径与行号、脚本输出、测试或运行记录。仅写在说明里的能力不等于已证明。
- **资源数量本身不加分**:存在 `references/`、`scripts/`、状态机或 SubAgent 都不是加分项。
- **不适用的复杂机制不扣分**:按任务本身需要的复杂度评估。简单 Skill 可以获得高分,复杂 Skill 也可能因过度设计失分。
- **100 分不得四舍五入**:满分必须包含运行证据,且所有维度均满足满分锚点。
- **静态最高 85 分**:前 7 个维度检查设计;最后 15 分只认真实行为验证,存在测试目录不等于已验证。
- 不为凑等级调整分数,不强制分布,也不拿两个参考项目当标准答案。

## 评测模式

| 模式 | 何时使用 | 证据范围 |
|---|---|---|
| `static` | 默认;单个或批量快速扫描 | 当前文件、资源,以及解析、链接、语法、`--help` 等无副作用检查;最高 85 分 |
| `verified` | 有现成记录,或允许安全前向测试 | `static` + 真实触发、边界/失败、产物和验收证据 |

无法安全运行时保持 `static`,明确证据缺口,不为追求高分制造副作用。

## 工作流

1. **确定范围**:接受一个 `SKILL.md`、一个 Skill 目录或仓库目录;记录当前 commit 与工作树状态,避免把快照结论说成永久事实。
2. **采集事实**:运行:

   ```bash
   python3 "${CLAUDE_SKILL_DIR}/scripts/collect_skill_facts.py" <目标> --format json
   ```

   若运行时不提供 `CLAUDE_SKILL_DIR`,从当前 `SKILL.md` 所在目录解析脚本路径。脚本只提供客观事实,不代替评分。
3. **建立画像**:记录产物(回答/文件/状态)、副作用(无/本地/外部)、路径(单路径/多分支/长流程)和依赖(无/本地/外部)。先判断需要多少复杂度,再评分。
4. **加载评分表**:完整读取 [`references/rubric.md`](references/rubric.md),逐维度评分。不要预读其它被评 Skill 的所有 references;只读取会改变当前判断的文件。
…

Read the whole file at its exact version.

How to install

The exact contents, by fingerprint
mdr add wangjs-jacky/jacky-skills/skill-design-benchmark@sha256:c4f70759802098fb

The line at the top of this page pins the label git:20260711.dae29b8, 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_uzfiwylp3pfiek5q.svg)](https://markdownregistry.com/a/art_uzfiwylp3pfiek5q)

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

Versions

The registry has recorded 1 version of this file, committed on Jul 11, 2026.

versioncommittedcommitsizeaudit
git:20260711.dae29b8 latest2026-07-11 dae29b8 4,974 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 (4974 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

wangjs-jacky/jacky-skills · 15 stars · license MIT · pushed 2026-10-19 · branch main

API

GET https://markdownregistry.com/api/v1/artifacts/art_uzfiwylp3pfiek5q
GET https://markdownregistry.com/api/v1/resolve?ref=wangjs-jacky/jacky-skills/skill-design-benchmark
GET https://markdownregistry.com/api/v1/blob/c4f70759802098fbca3a51316eefb0ff48d228128ba881fbe9dd1ed32fc320d0

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 wangjs-jacky/jacky-skills

AGENTS.md AGENTS.md file
wangjs-jacky/jacky-skills · AGENTS.md
git:20260711.ccfb5ec · audit B · 15 stars
CLAUDE.md CLAUDE.md file
wangjs-jacky/jacky-skills · CLAUDE.md
git:20261015.c21ad69 · audit A · 15 stars
agent-pipeline-benchmark SKILL.md skill
wangjs-jacky/jacky-skills · archived/agent-pipeline-benchmark/SKILL.md · Benchmark Evaluator:运行测试、评估质量、生成评分报告。由 agent-pipeline 编排器调用,也可通过 /agent-pipeline-benchmark 独立触发。
git:20260711.0fcbb55 · audit A · 15 stars
agent-pipeline-generate SKILL.md skill
wangjs-jacky/jacky-skills · archived/agent-pipeline-generate/SKILL.md · Generator Agent:基于计划实现代码,遵循 TDD 红→绿流程。由 agent-pipeline 编排器调用,也可通过 /agent-pipeline-generate 独立触发。
git:20260405.06b2dd2 · audit A · 15 stars
agent-pipeline-plan SKILL.md skill
wangjs-jacky/jacky-skills · archived/agent-pipeline-plan/SKILL.md · Plan Agent:调研代码库 + 生成实施计划和验收标准。由 agent-pipeline 编排器调用,也可通过 /agent-pipeline-plan 独立触发。
git:20260405.06b2dd2 · audit A · 15 stars
agent-pipeline SKILL.md skill
wangjs-jacky/jacky-skills · archived/agent-pipeline/SKILL.md · 三 Agent 流水线编排器。协调 Plan → Generate → Score 三个子 Agent 的执行、状态管理和质量循环。触发于 /agent-pipeline 或"流水线"、"三Agent"。
git:20260711.0fcbb55 · audit A · 15 stars
long-running-agent SKILL.md skill
wangjs-jacky/jacky-skills · archived/long-running-agent/SKILL.md · Use when starting a new session for a cross-session development project, or when the project uses Memory Bank structure…
git:20260405.06b2dd2 · audit A · 15 stars
multi-agent SKILL.md skill
wangjs-jacky/jacky-skills · archived/multi-agent/SKILL.md · 多 Agent 协作,并行调用多个 AI 模型并合并最佳结果。适用于需要高质量答案、代码审查、风险对冲的场景。
git:20260405.06b2dd2 · audit A · 15 stars
tw-benchmark SKILL.md skill
wangjs-jacky/jacky-skills · archived/tw-benchmark/SKILL.md · 基准评测 Agent。对 Generator 产出的代码进行多维评估:测试通过率、TypeScript 检查、代码质量。由 /tw-benchmark 或 task-workflow 编排器触发。
git:20260711.0fcbb55 · audit A · 15 stars
tw-generator SKILL.md skill
wangjs-jacky/jacky-skills · archived/tw-generator/SKILL.md · 代码生成 Agent。根据 PLAN.md 逐任务执行代码修改。由 /tw-generator 或 task-workflow 编排器触发。
git:20260711.0fcbb55 · audit A · 15 stars
tw-planner SKILL.md skill
wangjs-jacky/jacky-skills · archived/tw-planner/SKILL.md · 任务规划 Agent。调研代码库、分析需求、生成执行计划。由 /tw-planner 或 task-workflow 编排器触发。
git:20260711.0fcbb55 · audit A · 15 stars
CLAUDE.md@harness CLAUDE.md file
wangjs-jacky/jacky-skills · harness/CLAUDE.md
git:20261015.c21ad69 · audit A · 15 stars

Every file in wangjs-jacky/jacky-skills

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