research-pipeline-runner

SKILL.md

Goal: let a user trigger a full pipeline with one natural-language request, while keeping the run auditable (Units + artifacts + checkpoints).

In --strict runs, several semantic C3/C4 artifacts are treated as scaffolds until explicitly marked refined. This is intentional: it prevents bootstrap JSONL from silently passing into C5 writing (a major source of hollow/templated prose).

Create these markers only after you have manually refined/spot-checked the artifacts:

使用工作区+检查点端到端运行此存储库的单元+检查点研究管道（调查/综述/评论/调研/教程/系统综述/审稿）。 **触发**：运行管道、启动、继续执行、自动运行、写文章、调查/综述/评论/调研/教程/系统综述/审稿。 **使用时机**：用户希望端到端运行流程（创建`workspaces/<name>/`、生成/执行`UNITS.csv`、遇到HUMAN检查点停止等待）。 **如果**则跳过：用户明确要手动逐条执行（用`unit-executor`），或者你不应该自动推进到散文阶段。 **网络**：取决于所选管道（arXiv/PDF/引文验证可能需要网络；支持离线导入（如果可用）。 **Guardrail**：必须尊重检查点（无批准不写散文）；遇到 HUMAN 单元必须停止等待；禁止在 repo root 创建工作空间工件。 来源：willoscar/research-units-pipeline-skills。

查看原文