ccf-experiment-designer

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

CCF Experiment Designer

Invocation Controls

调用控制

CCFA Handoff Mode: PARTIAL (Recommended). Follow

metadata.ccf_skill_controls.handoff_question_mode

../ccf-common/references/handoff-modes.md

, and

../ccf-common/references/task-modes.md

Run

ccf-humanization

as the first publication-facing experiment preflight only when this skill produces or revises publication-facing prose, final manuscript tables/captions, or a publication method description. Do not load it for raw protocol planning, dataset/baseline/metric selection, execution queues, or evidence-schema design. When it applies, load

../ccf-humanization/references/experiment-discipline.md

, minimize smoke tests to unique changed critical paths, and allow only internally confirmed full method versions in manuscript text, final tables, captions, and claimed comparisons. In publication-facing wording, name the method and scientifically relevant configuration naturally without exposing confirmation, approval, or readiness status. Put version conflicts or necessary exceptions in a separate user-review warning; do not modify experiment or manuscript files merely to encode the warning.

CCFA 交接模式：PARTIAL（推荐）。遵循

metadata.ccf_skill_controls.handoff_question_mode

、

../ccf-common/references/handoff-modes.md

和

../ccf-common/references/task-modes.md

。

仅当本技能生成或修改面向出版物的实验相关文稿、最终手稿表格/说明文字，或出版物方法描述时，才需首先运行

ccf-humanization

作为面向出版物的实验预检。在进行原始方案规划、数据集/基线/指标选择、执行队列或证据schema设计时，请勿加载该技能。当适用时，加载

../ccf-humanization/references/experiment-discipline.md

，将冒烟测试最小化至唯一变更的关键路径，且仅允许内部确认的完整方法版本出现在手稿文本、最终表格、说明文字和公开的对比内容中。在面向出版物的表述中，自然地命名方法和科学相关配置，无需披露确认、审批或就绪状态。将版本冲突或必要例外情况放入单独的用户审核警告中；请勿仅为编码警告而修改实验或手稿文件。

Core Rule

核心规则

Design the smallest sufficient experiment package that tests the paper's central claims with complete method configurations verified by the internal gate. Build result tables and evidence-bound figure specs only from supplied real values or explicit placeholders. Never fabricate numbers, improvements, significance, benchmark ranks, or user-study outcomes. Do not expand protocols with repetitive smoke tests or implausible defensive cases. Publication-grade layout, palette, caption placement, and render QA belong to

ccf-visual-composer

. Follow the user's requested output shape: experiment plan, table, LaTeX table, figure spec, ablation list, or execution queue.

设计最小且充分的实验包，以测试论文的核心论点，且实验包需包含经内部审核验证的完整方法配置。仅从提供的真实值或明确占位符构建结果表和绑定证据的图表规格。切勿编造数值、性能提升、显著性、基准排名或用户研究结果。请勿用重复的冒烟测试或不合理的防御性案例扩展实验方案。出版物级别的布局、调色板、说明文字位置和渲染质量检查属于

ccf-visual-composer

的职责范围。遵循用户要求的输出形式：实验计划、表格、LaTeX表格、图表规格、消融实验列表或执行队列。

Modes

模式

```
design
```
: datasets, baselines, metrics, ablations, robustness, efficiency, failure analysis, and execution priority.
```
result-template
```
: fill-in tables with
```
TBD
```
placeholders.
```
result-presentation
```
: result tables, figure evidence plans, chart specs, caption facts, and missing-value markers from supplied real results.

```
design
```
：datasets、baselines、metrics、ablations、robustness、效率、故障分析及执行优先级。
```
result-template
```
：带有
```
TBD
```
占位符的可填充表格。
```
result-presentation
```
：基于提供的真实结果生成的结果表、图表证据计划、图表规格、说明文字事实及缺失值标记。

Workflow

工作流程

Run the
```
ccf-humanization
```
experiment preflight, then identify target venue, paper type, central claims, available results, confirmed method identities, and whether the task is planning or presenting results.
Extract the storyline from the idea or draft. Use
```
../ccf-paper-writer/references/storyline-blueprint.md
```
only as a schema, not as a writing handoff.
Map every major claim to sufficient evidence, dataset/workload, confirmed baseline, metric, and mechanism-relevant ablation. Add robustness or failure tests only when observed, plausible, claim-relevant, or venue-required; do not enumerate remote defensive cases.
If datasets or baselines are unknown, use public-safe search or hand off to
```
ccf-literature-searcher
```
; mark uncertainty instead of guessing.

Load

references/evidence-design.md

for venue-family expectations and

references/result-templates.md

for result tables.

For result presentation, preserve units, seeds, confidence intervals, dataset names, metric direction, and confirmed method version/configuration. Mark missing values explicitly; never fill them with simplified runs.
Retain only non-duplicative smoke tests for changed executable critical paths. Keep them outside publication evidence and do not use them as substitutes for full experiments.
Hand off to
```
ccf-visual-composer
```
for publication-grade figure/table layout, palettes, panel maps, captions, manuscript integration, and render QA.
Hand off to
```
ccf-paper-writer
```
for manuscript prose,
```
ccf-integrity-auditor
```
for number/claim consistency, and
```
ccf-submission-checker
```
for package or artifact readiness.

运行
```
ccf-humanization
```
实验预检，然后确定目标会议/期刊、论文类型、核心论点、可用结果、已确认的方法标识，以及任务是规划实验还是展示结果。
从想法或草稿中提取故事线。仅将
```
../ccf-paper-writer/references/storyline-blueprint.md
```
用作schema，而非写作交接工具。
将每个主要论点映射到充分的证据、数据集/工作负载、已确认的baseline、metric，以及与机制相关的ablation。仅当观察到、合理、与论点相关或会议/期刊要求时，才添加robustness或故障测试；请勿列举无关的防御性案例。
如果数据集或baselines未知，使用公开安全搜索或交接给
```
ccf-literature-searcher
```
；标记不确定性而非猜测。
加载
```
references/evidence-design.md
```
以了解会议/期刊领域的预期，加载
```
references/result-templates.md
```
以获取结果表模板。
在结果展示阶段，保留单位、随机种子、置信区间、数据集名称、metric方向，以及已确认的方法版本/配置。明确标记缺失值；切勿用简化实验填充。
仅保留针对变更后的可执行关键路径的非重复冒烟测试。将这些测试放在出版物证据之外，且勿将其作为完整实验的替代品。
交接给
```
ccf-visual-composer
```
处理出版物级别的图表/表格布局、调色板、面板映射、说明文字、手稿整合及渲染质量检查。
交接给
```
ccf-paper-writer
```
处理手稿文稿，交接给
```
ccf-integrity-auditor
```
处理数值/论点一致性检查，交接给
```
ccf-submission-checker
```
处理包或成果就绪性检查。

Adaptive Output Contract

自适应输出约定

Return the requested artifact first. For a result table request, output the table. For a figure request, output the evidence-bound figure spec and caption facts, then name

ccf-visual-composer

as next owner for visual composition when needed. For a full experiment-design request, use this default structure:

text

Mode:
Venue and assumptions:
Claim-evidence matrix:
Dataset / benchmark needs:
Confirmed method / baseline versions:
Baseline matrix:
Main experiments:
Ablations:
Robustness / failure / efficiency:
Smoke scope and deduplication:
Result tables or figure specs:
Missing values:
Execution priority:
No-fabrication status:
Next CCFA owner:

首先返回请求的成果。如果是结果表请求，输出表格；如果是图表请求，输出绑定证据的图表规格和说明文字事实，然后在需要时指定

ccf-visual-composer

作为下一环节负责人进行视觉创作。如果是完整实验设计请求，使用以下默认结构：

text

Mode:
Venue and assumptions:
Claim-evidence matrix:
Dataset / benchmark needs:
Confirmed method / baseline versions:
Baseline matrix:
Main experiments:
Ablations:
Robustness / failure / efficiency:
Smoke scope and deduplication:
Result tables or figure specs:
Missing values:
Execution priority:
No-fabrication status:
Next CCFA owner:

References

参考资料

```
references/evidence-design.md
```
: experiment and benchmark design.
```
references/result-templates.md
```
: fill-in result tables and presentation scaffolds.
```
../ccf-humanization/references/experiment-discipline.md
```
: confirmed full method gate, simplified-version prohibition, smoke-test scope, and experiment-to-paper checks.
```
../ccf-humanization/references/humanization-policy.md
```
: warning-only, non-injection, and defensive-case removal policy.

```
references/evidence-design.md
```
：实验与基准设计。
```
references/result-templates.md
```
：可填充结果表及展示框架。
```
../ccf-humanization/references/experiment-discipline.md
```
：已确认的完整方法审核、简化版本禁用、冒烟测试范围及实验到论文的检查规则。
```
../ccf-humanization/references/humanization-policy.md
```
：仅警告、不注入及移除防御性案例的政策。