skill-eval SKILL.md skillA
skill-eval from ninjasln-labs/agent-skills is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.
mdr add ninjasln-labs/agent-skills/skill-eval@v1.2.0Pins v1.2.0 in mdr.lock. Pin the fingerprint instead.
- Audit grade
- A, 17 of 17 checks passed
- Kind
- SKILL.md skill
- Repository
- ninjasln-labs/agent-skills
- Path
skill-eval/SKILL.md- Versions recorded
- 4
- Latest version
- v1.2.0, committed Sep 29, 2026
- First version recorded
- Sep 18, 2026
- Size
- 9,732 bytes
- Repository stars
- 3
- License
- none declared
Behaviorally evaluate a skill in two modes: pass-rate (define 3-5 representative tasks, run each N>=3 times with and without the skill loaded in fresh agents, compare pass rates) and trigger rate (build ~20 queries — 8-10 should-trigger plus 8-10 near-miss — run each 3 times, score per-query trigger rate). Both report a table plus concrete improvement items that feed back into the skill's SKILL.md. Use when a skill changes, at quarterly evaluation, when accepting a new skill, or when…
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 Eval（高频 skill 动态评估：行为 pass-rate ＋ 触发率）
## 定位
高频 process skill 的**动态评估**，两模式互补：① **行为 pass-rate**（有/无 skill 各跑 N 次 → 对比 → 失败案例反哺正文）；② **触发率 trigger rate**（描述在该触发时是否触发、不该触发时是否误触发）。与 `skill-description-audit`（静态描述合规）互补。
## 何时使用
- skill 变更后（行为是否真变好）
- 季度评估（基线节奏）
- 新 skill 验收（交付后先建基线）
- 描述触发不可靠时（应触发不触发／误触发）
## 输入
- 目标 skill（首批：code-review / project-handoff / writing-plans / stage-gate）
- 评估记录目录：`.scratch/neonforge-v1/skill-eval/<skill>-<YYYY-MM-DD>/`
## 流程
```
- [ ] 1. 定义代表任务：3-5 个（见下节）
- [ ] 2. 无 skill 基线：每任务跑 N 次（N≥3，fresh agent、无对话种子）
- [ ] 3. 有 skill 组：同任务同 prompt 跑 N 次（fresh agent + 加载目标 skill 指令）
- [ ] 4. 逐次判定 pass/fail（按任务 AC，机器可验证优先——文件存在/输出匹配/命令结果）
- [ ] 5. 汇总 pass-rate 对比表（任务 | 无 skill | 有 skill | Δ）
- [ ] 6. 失败案例分析：成功/失败样例各取代表，定位 skill 改进点
- [ ] 7. 产出报告 + 改进项写回 skill SKILL.md（或作为下轮验收项）
- [ ] 8. 结果入 SKILLS-MAP 或该 skill README（基线表）
```
## 代表任务定义（3-5 个）
每个任务 = 一段**独立 prompt** + 可判定的 AC：
| 要求 | 说明 |
|------|------|
| 覆盖核心能力 | 任务必须命中 skill 的标志性能力（如 stage-gate 的「跑 S{N} 门禁」） |
| 有代表性失败模式 | 选任务时预判「无 skill 会怎么错」（漏验证/散文断言/只验不修） |
| AC 机器可验证 | 优先：文件存在且含 X / 命令输出匹配 / 断言数正确；次选：结构化报告字段齐全 |
| 现实规模 | 用真实仓库的最小场景（同仓库 fixture），不用玩具示例 |
## 运行纪律
…Read the whole file at its exact version.
How to install
mdr add ninjasln-labs/agent-skills/skill-eval@sha256:1442664ac906d0dfThe line at the top of this page pins the label v1.2.0, 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.
[](https://markdownregistry.com/a/art_ofaolsevwem4mpzs)
1 badge image fetches in 30 days. What the badge shows.
Versions
The registry has recorded 4 versions of this file, between Sep 18, 2026 and Sep 29, 2026, about one every 4 days.
| version | committed | commit | size | audit | |
|---|---|---|---|---|---|
| v1.2.0 latest | 2026-09-29 | 0febb7a | 9,732 B | A | view · diff |
| v1.1.0 | 2026-09-28 | 1fd959c | 6,519 B | A | view · diff |
| v1.0.2 | 2026-09-22 | 75c75c0 | 4,636 B | A | view · diff |
| v1.0.0 | 2026-09-18 | cbc6719 | 4,461 B | A | view |
Audit of the latest version
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (9732 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
ninjasln-labs/agent-skills · 3 stars · license none declared · pushed 2026-10-05 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_ofaolsevwem4mpzs GET https://markdownregistry.com/api/v1/resolve?ref=ninjasln-labs/agent-skills/skill-eval GET https://markdownregistry.com/api/v1/blob/1442664ac906d0df301ecb9573124ceda8d363e26a0d42cecb78758a1b4e70a7
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 ninjasln-labs/agent-skills
Every file in ninjasln-labs/agent-skills