Individual skill
light-experiment
Execute and document computational research experiments, especially after light-design. Use for 实验落地、实现并运行实验、数值仿真、训练与消融、算法或系统基准测试、失败诊断、续跑、实验过程及原始产物归档. Connect frozen designs to actual code, inputs, bounded runs, logs, raw outputs, QC and traceable findings. Idea discovery, study design alone, physical laboratory operations and manuscript writing belong elsewhere.
light-experiment/SKILL.md
Source commit 6597430ad09d8f7c3518ec4cb08829be2b417e18
What was checked
- Matches the observed repository HEAD
- Yes
- Scanner rules are current
- Yes
- Repository scan complete
- No
- Skill scan complete
- No
- Observed behavior
- Network or broad shell
- Last checked
- Nov 8, 2026
These are observations at the recorded commit, not a guarantee of safety. Review the source before running it.
Install this skill
Choose an agent in your Stack or copy a command below. Commands fetch the pinned commit and copy only this skill directory. Existing destinations are refused. Copying a command does not install anything.
An install command is unavailable: current rules, complete repository and skill scans, an eligible behavior level and a permitted license are required.
Repository context
- Repository stars (not skill ratings)
- 883
- Repository last updated
- Nov 7, 2026
- License
- MIT
Observed agent integration: No specific integration evidence. Standard SKILL.md installation does not imply verified compatibility.
- academic-writing Repository topic
- ai-agent Repository topic
- claude-code Repository topic
- codex Repository topic
- llm-skills Repository topic
- paper-writing Repository topic
- research-tools Repository topic
- scientific-workflow Repository topic
- agent-skills Repository topic
- ai-skills Repository topic
- design Derived category
- operations Derived category
- writing Derived category
- research Derived category
Files and excerpt
File metadata and permitted, redacted excerpts come from the recorded scan. Source links are pinned to that commit.
Browse files
.gitignoreasset · 56 bytesSKILL.mdcore · 13,811 bytesSKILL.mdcore · 13,811 bytesSKILL.mdcore · 13,811 bytesSKILL.mdcore · 13,811 bytesSKILL.mdcore · 13,811 bytesagents/openai.yamlasset · 332 bytesevaluations/2026-11-06-upgrade.mddoc · 5,393 bytesevaluations/scenarios.mddoc · 4,078 bytesevaluations/validation-report.mddoc · 4,680 bytesevaluations/validation-results.jsonasset · 2,150 bytesreferences/artifact-contract.mddoc · 12,950 bytesreferences/domains.mddoc · 4,860 bytesreferences/evidence-and-handoff.mddoc · 4,528 bytesreferences/execution.mddoc · 5,063 bytesreferences/pipeline.mddoc · 4,348 bytesreferences/project-layout.mddoc · 6,114 bytesreferences/provenance.mddoc · 4,021 bytesreferences/repository-audit.jsonasset · 18,580 bytesreferences/sources.mddoc · 17,606 bytesscripts/experiment.pyruns · 50,950 bytestemplates/assessment.example.jsonasset · 385 bytestemplates/execution-report.mddoc · 1,578 bytestemplates/run-spec.example.jsonasset · 1,183 bytestests/test_execution_evidence.pyruns · 11,562 bytestests/test_experiment.pyruns · 25,635 bytestests/test_pipeline_integration.pyruns · 11,232 bytes
Read excerpt
# Light Experiment
## 项目目录与续作
需要落盘或继续已有项目时,先读 [项目目录约定](references/project-layout.md),按其中的本阶段位置、版本与关联规则保存;用户指定路径和旧项目优先。包内资源相对本 skill,研究产物相对真实项目,二者不要混用。只做局部讨论时无需初始化目录。
项目已有 `Memory/` 时先恢复当前记忆、检索本轮设计/运行/故障,再回原生设计、运行和日志核对版本与状态;记忆不能代替读取真实证据。项目已有 `Workflow/` 时接住当前任务、输入版本、依赖和剩余预算。缺少这些目录时直接处理实际材料,不为一次运行先创建整个科研流程。
把计算实验的设计**实际实现、运行、检查并留下可追溯的原始证据**。覆盖计算机、数学与理工科的计算研究,包括 ML、数值计算、仿真、算法和系统评测。默认中文交流。不执行湿实验、仪器操作或招募。
宿主负责实现、选择工具、科学判断和持续推进;附带的 Python 标准库工具负责冻结交接、限额运行、原样记录和完整性检查。它不是自动科研平台,不调用模型、不联网、不安装依赖。只整理已有结果时不启动新实验;用户要求执行时,在已明确的范围和预算内继续完成,不把每一步都变成确认问题。
## 按任务选入口