Skip to content
Agentmart
Submit a tool

experiment-selection-ops

Use when a technical choice needs a comparative experiment, including visual designs, animations, functions or libraries; when the decision report lacks…

Skill by wangjs-jacky

IndexedMIT★ 15
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/wangjs-jacky-experiment-selection-ops

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Use when a technical choice needs a comparative experiment, including visual designs, animations, functions or libraries; when the decision report lacks traceable evidence; or when articles, videos and cases should inform such experiments. 适用于技术选型、对比实验与方法积累,不用于只查API或只制作单一成品。

Use cases: documents, backend

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/experiment-selection-ops

What's inside

Skill name
experiment-selection-ops
Description
Use when a technical choice needs a comparative experiment, including visual designs, animations, functions or libraries; when the decision report lacks traceable evidence; or when articles, videos and cases should inform such experiments. 适用于技术选型、对比实验与方法积累,不用于只查API或只制作单一成品。
License
MIT
Instructions
32 lines in SKILL.md · view on GitHub

Files (43)

  • .gitignore
  • SKILL.md
  • THIRD_PARTY_NOTICES.md
  • openai.yaml · agents/
  • decision-report.template.md · assets/
  • run-plan.template.md · assets/
  • claude-scholar-MIT.txt · licenses/
  • k-dense-MIT.txt · licenses/
  • squares-LICENSE.txt · licenses/
  • global-agents-prompt.md · references/
  • knowledge-schema.md · references/
  • blocking-and-interactions.md · references/methods/
  • design-and-execution.md · references/methods/
  • diagnostic-figures.md · references/methods/
  • index.md · references/methods/
  • materials.md · references/methods/
  • question-to-contrast.md · references/methods/
  • reporting.md · references/methods/
  • statistics.md · references/methods/
  • ops-model.md · references/
  • catalog.json · references/sources/
  • kd-discriminating-tests.txt · references/sources/excerpts/
  • kd-experimental-design.txt · references/sources/excerpts/
  • kd-factorial-doe.txt · references/sources/excerpts/
  • kd-hypothesis-quality.txt · references/sources/excerpts/
  • kd-randomization-blocking.txt · references/sources/excerpts/
  • kd-statistical-analysis.txt · references/sources/excerpts/
  • scholar-analysis-pitfalls.txt · references/sources/excerpts/
  • scholar-evidence-propagation.txt · references/sources/excerpts/
  • scholar-figure-catalog.txt · references/sources/excerpts/
  • scholar-question-card.txt · references/sources/excerpts/
  • scholar-results-analysis.txt · references/sources/excerpts/
  • scholar-results-report.txt · references/sources/excerpts/
  • squares-experiment-loop.txt · references/sources/excerpts/
  • index.md · references/sources/
  • index.md · references/task-types/animation/
  • index.md · references/task-types/function-library/
  • compare-csv-import.md · references/task-types/function-library/operations/
  • compare-node-config-validation.md · references/task-types/function-library/operations/
  • index.md · references/task-types/
  • index.md · references/task-types/visual-design/
  • compare-docs-homepages.md · references/task-types/visual-design/operations/
  • version.json

SKILL.md (first lines)

# 选型实验 OPS

以**任务类型 → 场景/比较目标 operations**积累方法。工具、库、模型都是候选。帮助形成有直接证据的条件建议,允许无定论;一次案例的赢家不等于实验框架的赢家。

安装/配置时先读[可选全局前置提示词](references/global-agents-prompt.md)。普通使用直接开始;保存提示词不会修改全局规则。

## 开始与路由

1. 从本 Skill 实际目录确认唯一知识根。历史只读快照只用于指定测试,写回遵循[知识协议](references/knowledge-schema.md),不能成为第二个当前库。
2. 读[类型索引](references/task-types/index.md),只加载命中类型的档案和目标 operation。分类不清或准备改变分类轴时再读[分类模型](references/ops-model.md)。没有经验就按当前条件设计,明确尚未验证。
3. 复核当前用户目标、授权、版本、工具、环境、预算与旧经验适用范围。模型与 effort 记录本次请求及可取得的有效字段,不固定未来型号。只读调研与真实运行分别说明。

## 每轮闭环

- **读方法、复核、设计执行**:按[设计与真实执行](references/methods/design-and-execution.md),在取结果前留下问题、候选、案例、变量、判断及停止条件;小实验可以把短计划放同一文件,不必拆成多个研发任务。
- **检查与解释**:按类型取得可打开/可播放/可运行的相应产物,保留失败、缺失与偏离。有连续数据推断需求才读[统计方法](references/methods/statistics.md)。
- **报告**:用[报告指南](references/methods/reporting.md)串起必要过程、数字、解释与条件建议;关键错误阻断依赖它的主张。图表、文字和案例按决策需要选用,无统一图数、总分或篇幅。
- **写回、更新索引、交付**:每次完成、失败或中断都按[知识协议](references/knowledge-schema.md)判断并执行;报告实际写回位置或未保存原因。自动沉淀是 Agent 工作协议,不是后台服务或宿主钩子。

资料输入是文章、视频/转录或优秀案例时,读[方法提炼](references/methods/materials.md);同时提取选题、比较过程、变量、测试办法与呈现,不只摘取写作风格。

## 证据与执行边界

0.1.0初始知识来自固定上游与旧原版观察。08实际用0.1.0取得两类同题局部证据,随后一次修订和受影响项复验结束;09已合并两份知识补丁的最终可采纳内容,两个operation为partial,仍无完整成功日期。历史单调总耗时invalid、初版协调读取顺序偏离、两视觉子CLI超时、残留轻微卡点与未测/真人未知均保留。0.1.1时钟修订仅有受影响项检查;09在0.1.2 / 0.1.2-evidence合并后,又从01实际保存原文新增三项documented共享方法及薄路由,当时最终为0.2.0 / 0.2.0-methods,该阶段未重跑业务。任务10随后实际使用该0.2版本完成CSV场景;三初判冻结后的直接证据支持三共享方法在这个场景的局部应用。现合入CSV operation、路由和相关边界,实现仍为0.3.0,知识现为0.3.1-final-evidence;保持partial且无成功日期。新operation不是原0.2读取内容。2026-10-17三路fresh独立受影响项复验结束:数据角色实际从新入口检索CSV并有限应用,核对移动离线安装、正常/缺worker同出口和共享源身份;逻辑与可读性角色支持受影响说明,未代验数据。0.3未作业务整轮,跨类型、方法因果优势及真人理解仍未知。本次最终变化只有证据/日期/阶段事实,无新增步骤或路由。[版本说明](version.json)与[来源索引](references/sources/index.md)保留阶段范围;来源摘录不是额外Skill入口,任务完成不改变知识等级。

使用宿主已授权的执行工具。网页操作遵守其治理与执行技能;环境要求 Ego 时先读 ego-ops 再用 ego-browser。缺失工具则披露相应未验项。专业领域约束仍由其权威规则负责,本 Skill 不复制工具手册或扩大安装、发布、付费与外部写入权限。

真人偏好和理解只能由真人反馈。按需的独立多 Agent 初判先各自冻结,再查直接证据协调,不能多数票抵消关键事实错误;没有独立审阅时如实说明。

调用例:用本 Skill 比较同一文档首页的两种设计方法,交付真实页面、中性评看入口与条件建议,并回写这类比较经验。此例不自动启动实验。

From wangjs-jacky/jacky-skills, MIT.

More from wangjs-jacky/jacky-skills

Signals

Stars
15
Forks
4
Open issues
2
Last commit (this folder)
Oct 19, 2026 (9 days ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from wangjs-jacky

Cite this page

experiment-selection-ops skill. Agent Mart. https://agentmart.dev/skills/wangjs-jacky-experiment-selection-ops (last updated 2026-10-24)