Paper digger plan

Skill Iu1ky/paper-digger/plugins/paper-digger/skills/paper-digger-plan

Use when converting a confirmed idea and venue into falsifiable hypotheses, an experiment matrix, dependencies, and one kill-early test. Triggers on 研究计划, 实验矩阵, experiment plan, 最小验证.From its SKILL.md

Install

npx -y skills add Iu1ky/paper-digger --skill paper-digger-plan

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

16 days oldThe repository was created 16 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.3 KB, 738 tokens by cl100k_base, as published. Nobody here has run it

Paper Digger Plan — 安排计划(Phase 4)

输入 confirmed idea(01_ideation/)+ confirmed venue(02_venue/)+ Phase 3 文献 gap map,输出一份研究计划:假设 → 实验矩阵 → 依赖图 → 里程碑,并指定一个最小验证实验(kill-early)。

运行时

定位同级 paper-digger skill，把它的 scripts/ 加入 Python import path 后调用 paper_digger.plan.*。不要求系统已安装全局 pd 包。

实验矩阵

每个实验是一张 dict: {id, question, mode: dry|wet|theory, method, success_criteria, deps: [id], est_cost, is_min_validation}

mode:dry(代码/仿真/数据分析,可真跑)· wet(物理/湿/人类受试,出 protocol 由人执行)· theory(推导)。
success_criteria:这个实验「算成功」的明确判据(避免事后挪动靶子)。
deps:依赖的实验 id(决定执行顺序与可并行集)。
is_min_validation:恰好一个实验标 True —— 最便宜、最能证伪核心主张的那个,先跑它,触发评价节点②(kill-early)。

流程

从 idea/venue/gap 提炼可证伪的假设/主张。
为每个假设设计实验,填实验矩阵;指定一个 is_min_validation。
校验 + 排序 + 落盘:paper_digger.plan.save_plan(workspace, hypotheses, experiments) —— 内部 validate_matrix(模式合法、id 唯一、deps 存在、恰好一个最小验证)+ dependency_order(拓扑序,有环报错),写 04_plan/research_plan.md。
checkpoint:呈现假设 + 矩阵 + 执行顺序 + 预估成本,由人批准。
落定:paper_digger.plan.confirm_plan(workspace, experiments) —— 把矩阵种入 state.experiments(status=planned)、置 state.decisions.plan_approved,供 paper-digger-experiment 执行。

铁律

先有判据再做实验:success_criteria 必须在执行前定死。
恰好一个最小验证:它是 kill-early 的触发点;没有就不算计划完成。
不臆造可行性:est_cost/mode 要诚实(算力/湿实验/时间约束见 00_profile/)。

What ships with it: 1 file

250 B alongside SKILL.md

agents/

openai.yaml250 B