菜单
Build a lightweight task-visible data and resource profile for AutoSciRub. Use after rubric skeleton and literature grounding when Codex needs to inspect available files, datasets, code, schemas, units, labels, tools, constraints, and goal feasibility before synthesizing executable criteria.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/zjunlp/AutoSciRub --skill task-data-exploration命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
正在显示 SKILL.md
| name | task-data-exploration |
| description | Build a lightweight task-visible data and resource profile for AutoSciRub. Use after rubric skeleton and literature grounding when Codex needs to inspect available files, datasets, code, schemas, units, labels, tools, constraints, and goal feasibility before synthesizing executable criteria. |
Profile the task-visible environment without running full experiments. This implements phi_data(E_i) from AutoSciRub.
The profile determines which literature-suggested analyses are feasible with the current files, tools, labels, and compute context.
All .autoscirub/ paths below refer to the state directory resolved by the controller. When invoked alone, use the user’s override, then AUTOSCIRUB_STATE_DIR, then project config state_dir, then .autoscirub/.
Read:
.autoscirub/rubric_skeleton.json.autoscirub/literature_grounding.jsonRespect exclusions from the user or task harness. Do not inspect hidden answers, target studies, private grading files, or files outside the task-visible environment.
Use lightweight inspection first:
Avoid full model training, expensive simulation, final figure generation, or report writing in this step.
Write .autoscirub/task_data_profile.json:
{
"schema_version": "1.0",
"files": [
{
"path": "relative/path",
"format": "csv | json | hdf5 | image | text | code | other",
"size_or_shape": "...",
"key_fields": ["..."],
"role": "input | label | reference | metadata | geometry | time_series | model_output | code | other",
"notes": "..."
}
],
"datasets": [
{
"name": "...",
"paths": ["..."],
"entities": ["..."],
"measurements": ["..."],
"labels_or_targets": ["..."],
"constraints": ["..."]
}
],
"relationships": ["join, alignment, temporal, spatial, or hierarchy relation"],
"constraints": ["missing labels, limited samples, compute limits, unavailable reference values"],
"goal_support": [
{
"goal_id": "G1",
"supporting_files": ["..."],
"feasible_analyses": ["..."],
"unsupported_requirements": ["..."],
"notes": "..."
}
],
"dependency_notes": ["libraries or external tools that may be needed, without activation commands"]
}