跳到正文
cccskills
免费在 GitHub 公开

ieee-experiment

Design, audit, and strengthen experimental validation for IEEE manuscripts using routed claim-evidence checks. Use when planning experiments, selecting baselines, designing ablations, choosing metrics, checking fairness, adding robustness tests, explaining results, analyzing complexity, building claim-to-evidence matrices, or responding to reviewer concerns about insufficient experiments.

中文简介正在准备中。当前显示原文说明。

查看安装方法

包含的文件(33)

  • SKILL.md2.3 KB
  • agents/openai.yaml226 B
  • manifest.yaml3.3 KB
  • static/core/claim-evidence-matrix.md599 B
  • static/core/output-format.md465 B
  • static/core/stance.md391 B
  • static/core/workflow.md632 B
  • static/fragments/evidence_type/ablation.md266 B
  • static/fragments/evidence_type/baseline.md375 B
  • static/fragments/evidence_type/complexity.md368 B
  • static/fragments/evidence_type/real-world.md368 B
  • static/fragments/evidence_type/reproducibility.md372 B
  • static/fragments/evidence_type/robustness.md326 B
  • static/fragments/evidence_type/statistical.md349 B
  • static/fragments/failure_mode/insufficient-reproducibility.md279 B
  • static/fragments/failure_mode/missing-traditional-baseline.md300 B
  • static/fragments/failure_mode/no-condition-test.md377 B
  • static/fragments/failure_mode/overclaimed-results.md340 B
  • static/fragments/failure_mode/unfair-comparison.md320 B
  • static/fragments/failure_mode/weak-ablation.md275 B
  • static/fragments/stage/audit.md149 B
  • static/fragments/stage/planning.md213 B
  • static/fragments/stage/result-writing.md210 B
  • static/fragments/stage/reviewer-response.md292 B
  • static/fragments/task_type/classification.md466 B
  • static/fragments/task_type/communications.md409 B
  • static/fragments/task_type/control.md369 B
  • static/fragments/task_type/detection.md401 B
  • static/fragments/task_type/general.md257 B
  • static/fragments/task_type/hardware-system.md382 B
  • static/fragments/task_type/optimization.md348 B
  • static/fragments/task_type/regression.md335 B
  • static/fragments/task_type/signal-processing.md373 B

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

IEEE Experiment Router

Use this skill to decide whether the experiments prove the paper's claims. The primary output is an evidence audit, not generic advice.

Do not design experiments from memory alone. Follow the routing protocol and load the selected fragments.

Routing Protocol

  1. Read manifest.yaml.
  2. Read every file listed under always_load.
  3. Detect the axes:
    • task_type: classification / detection / regression / control / signal-processing / communications / optimization / hardware-system / general.
    • evidence_type: baseline / ablation / robustness / complexity / statistical / real-world / reproducibility.
    • failure_mode: missing-traditional-baseline / unfair-comparison / weak-ablation / no-condition-test / overclaimed-results / insufficient-reproducibility.
    • stage: planning / audit / result-writing / reviewer-response.
  4. State the detected axes in one short line.
  5. Load only the matching fragments.
  6. Build or update a claim-evidence matrix.
  7. Identify missing experiments by reviewer impact.

Output Contract

Default output:

Detected axes: task_type=..., evidence_type=..., failure_mode=..., stage=...

Claim-evidence matrix
Claim | Required evidence | Current evidence | Missing experiment | Review risk

Priority fixes
1. ...

For experiment planning, return an experiment plan with baselines, metrics, variables, controlled conditions, and expected claims.

For result writing, return IEEE-style result paragraphs and flag any claim that lacks evidence.

Red Lines

Do not invent numerical results, datasets, baseline performance, p-values, hardware metrics, or statistical significance.

Do not recommend unnecessary experiments that do not support a stated claim.

Do not treat "more experiments" as automatically better. Prioritize experiments that close the reviewer's proof gap.

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

Check, clean, verify, and improve IEEE citations, references, BibTeX entries, citation positioning, related-work logic, DOI metadata, venue names, and reference formatting using routed citation and metadata guidance. Use when preparing IEEE manuscripts, fixing bibliography errors, verifying reference completeness, improving related work, checking whether citations support claims, or identifying missing classic/recent baselines.

中文简介正在准备中。当前显示原文说明。

CloudWave818/ieee-skills3942026年8月12日 更新

Audit, redesign, generate, and improve IEEE manuscript figures, tables, captions, result presentation, plotting scripts, visual polish, hybrid Python/R plus vector-editor workflows, publication-ready matplotlib house style, and layout using routed visual-evidence checks. Use when checking or creating plot readability, axes, legends, line overlap, table structure, metric reporting, captions, grayscale accessibility, IEEE one-column/two-column layout, figure/table placement, claim-evidence mapping, reviewer first-impression risks, matplotlib/seaborn or ggplot2 figure code, SVG/PDF/TIFF/PNG export, Illustrator/Inkscape/PowerPoint/Figma finishing plans, or IEEE-style engineering evidence graphics.

中文简介正在准备中。当前显示原文说明。

CloudWave818/ieee-skills3942026年8月12日 更新

Diagnose, fix, and improve IEEEtran LaTeX manuscripts using routed template, layout, float, equation, algorithm, table, bibliography, and PDF-compliance guidance. Use when fixing IEEE LaTeX compile errors, overfull boxes, float placement, two-column layout, figure sizing, tables, equations, algorithms, BibTeX, citation commands, template compliance, or final submission PDF checks.

中文简介正在准备中。当前显示原文说明。

CloudWave818/ieee-skills3942026年8月12日 更新

Read, decompose, and analyze IEEE-style papers using routed technical-reading workflows. Use when extracting problem setting, object-method-condition logic, contributions, method pipeline, equations, datasets, baselines, metrics, ablations, limitations, replication details, citation role, related-work positioning, implementation notes, or reviewer-style assessment from a paper, PDF, abstract, or manuscript excerpt.

中文简介正在准备中。当前显示原文说明。

CloudWave818/ieee-skills3942026年8月12日 更新

Polish, translate, diagnose, and restructure technical manuscript text into IEEE-style academic English using routed failure-mode guidance. Use for IEEE abstracts, introductions, related work, methods, experiments, conclusions, figure captions, rebuttals, Chinese-to-English technical writing, AI-like or template-like academic prose, and engineering prose that needs clearer problem logic, contribution framing, method rationale, evidence discipline, natural author voice, or reviewer-ready wording.

中文简介正在准备中。当前显示原文说明。

CloudWave818/ieee-skills3942026年8月12日 更新

Draft, diagnose, and improve IEEE rebuttals, revision plans, cover letters, response letters, and point-by-point replies using routed reviewer-comment handling. Use when responding to IEEE reviewer or editor comments about novelty, scope, method, experiments, baselines, figures, tables, citations, writing, LaTeX formatting, limitations, compliance, technical correctness, or reviewer comments that appear generic, contradictory, unsupported, careless, or possibly AI-generated.

中文简介正在准备中。当前显示原文说明。

CloudWave818/ieee-skills3942026年8月12日 更新

查看 CloudWave818 的全部技能

报告此技能的问题