ai-safety-testing SKILL.md skillA
ai-safety-testing from naodeng/awesome-qa-skills is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.
mdr add naodeng/awesome-qa-skills/ai-safety-testing@v1.0.0Pins v1.0.0 in mdr.lock. Pin the fingerprint instead.
- Audit grade
- A, 17 of 17 checks passed
- Kind
- SKILL.md skill
- Repository
- naodeng/awesome-qa-skills
- Path
skills/zh/testing-types/ai-safety-testing/SKILL.md- Versions recorded
- 3
- Latest version
- v1.0.0, committed Oct 29, 2026
- First version recorded
- Sep 15, 2026
- Size
- 4,915 bytes
- Repository stars
- 301
- License
- a license file GitHub does not identify
Use this skill when you need evidence-bounded safety policy, abuse categories, refusal/redirect behavior, privacy, escalation, and Human risk decisions; triggers include AI 安全 and AI safety.
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
# AI 安全测试（中文版） ## 何时使用 - 需要围绕安全政策、滥用类别、拒答/重定向、隐私、升级与 Human 风险决策开展有证据的分析、设计或验证准备。 - 需要评审 Agent、RAG 或 LLM 方案、结果或证据，并形成可执行改进项。 - 输入不完整，但仍需交付带假设、信息缺口和人工决策边界的受限初版。 ## 输出格式选项 - 默认输出 Markdown，按领域风险、证据状态、优先级和边界组织。 - 用户要求表格、CSV、JSON 或工单格式时，保留相同的发现字段、证据和决策边界。 - 输出进入自动化流程前，先确认 schema、枚举值、必填项和证据来源。 ## 如何使用 1. 先阅读并遵循 `prompts/ai-safety-testing.md`，包括输入审计、领域覆盖和输出顺序。 2. 提取范围、环境、版本、时间窗口、约束、成功标准和可用证据，重点检查安全政策、滥用类别、拒答与重定向、隐私边界、升级决策。 3. 先区分已确认事实、证据支持的推断、候选建议和 Human 决策，再按风险与证据强度排序。 4. 把高风险项转成前置条件、步骤、预期结果或判断标准、所需证据和验证方法。 5. 信息不足时交付受限初版，明确哪些结论不能成立，不把静态材料写成执行结果。 ## 参考文件 - 校准领域结果、正确对照或证据缺口时，读取 [references/domain-decision-example.md](references/domain-decision-example.md) 的脱敏输入、逐项预期与复算依据。 - 每次执行必须读取 `prompts/ai-safety-testing.md`；它是本 Skill 的完整执行规范。 - 需要评测时读取 `evals/eval.yaml` 和 `evals/cases/` 下的匹配用例。 - 只有目录实际存在且任务需要时，才读取 `references/`、`examples/`、`scripts/` 或 `output-formats.md`，不要假设不存在的资产。 ## 领域执行要点 1. 固定政策 ID/版本/适用对象，建立 类别/条件→允许/限制/拒答/升级→可提供内容/禁止副作用→证据 矩阵；缺政策不自行认定安全批准。 2. 用政策描述的危险样本与近邻正常请求配对；分别检查危险处理正确性、允许请求完成质量和过度拒答，不能只汇总拒答率。 3. 限制/重定向核对实际内容及工具动作，不能只看拒答开头；升级记录触发原因、责任角色及待 Human 决策状态。 4. 标签记录政策来源、样本版本、预期分支与实际分支；未提供执行输出保持 NOT_RUN，未知政策边界保持 UNKNOWN。 ## 核心约束 …
Read the whole file at its exact version.
How to install
mdr add naodeng/awesome-qa-skills/ai-safety-testing@sha256:534b093dcfadc711The line at the top of this page pins the label v1.0.0, 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_lw25z72ts5bx7e5h)
2 badge image fetches in 30 days. What the badge shows.
Versions
The registry has recorded 3 versions of this file, between Sep 15, 2026 and Oct 29, 2026, about one every 22 days.
| version | committed | commit | size | audit | |
|---|---|---|---|---|---|
| v1.0.0 latest | 2026-10-29 | 1d0d58f | 4,915 B | A | view · diff |
| git:20261011.7d9502b | 2026-10-11 | 7d9502b | 4,606 B | A | view · diff |
| git:20260915.27b248b | 2026-09-15 | 27b248b | 3,738 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 (4915 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
naodeng/awesome-qa-skills · 301 stars · license a license file GitHub does not identify · pushed 2026-11-06 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_lw25z72ts5bx7e5h GET https://markdownregistry.com/api/v1/resolve?ref=naodeng/awesome-qa-skills/ai-safety-testing GET https://markdownregistry.com/api/v1/blob/534b093dcfadc71179efb0c9e52ce8fd0d4437e45521e8140fc05ac131fdaad3
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 naodeng/awesome-qa-skills
Every file in naodeng/awesome-qa-skills