paper-reading-workbench

name	paper-reading-workbench
description	Organize one or more researcher-provided PDFs, DOI, arXiv identifiers, or existing paper records into prioritized reading cards, section-level answers, and a cross-paper learning summary. Use when a researcher wants a practical reading workspace rather than a single PDF note card.
tools	["read_file","write_file","write_structured_file","list_files","fetch_paper_metadata","openalex_search","semantic_scholar_search","crossref_search","fetch_paper_pdf","extract_pdf_text","extract_paper_sections","save_paper_note","ask_human","update_skill_workflow","finish_task"]
strict_tools	true
model_tier	standard
temperature	0.15
allowed_read_prefixes	["user_inputs/paper-reading-workbench/","literature/","user_seeds/","_runtime/skill_sessions/"]
allowed_write_prefixes	["user_inputs/paper-reading-workbench/","literature/skill_reading_workbench/"]
outputs_expected	{"reading_index":"literature/skill_reading_workbench/reading_index.md","paper_cards":"literature/skill_reading_workbench/paper_cards.json","question_answers":"literature/skill_reading_workbench/question_answers.md","learning_summary":"literature/skill_reading_workbench/cross_paper_summary.md","workflow_manifest":"literature/skill_reading_workbench/manifest.json"}
interaction	{"mode":"guided","language":"zh-CN","summary":"上传/列出多篇 PDF、DOI、arXiv 或已有记录，按问题自动建立阅读队列、来源锚定卡片与跨论文学习总结。","request_required":true,"request_prompt":"请说明要读哪些论文、希望回答的研究问题、优先轴和是否接受系统先解析 DOI/PDF 后再问补料。","example_request":"阅读这组 PDF，比较它们的 memory mechanism、实验边界和最值得复用的设计 rationale。","required_inputs":[{"id":"reading_request","label":"阅读任务与问题","description":"说明论文标识或上传路径、要回答的问题、优先级和希望的阅读深度。","paths":"[Truncated]","extensions":"[Truncated]","min_bytes":40,"example":"user_inputs/paper-reading-workbench/request.md"}],"optional_inputs":[{"id":"source_list","label":"PDF/DOI/标题/记录清单","description":"可选；一行一个 workspace 相对路径或稳定标识。也可在交互中逐项上传/补充。","paths":"[Truncated]","extensions":"[Truncated]","min_bytes":10,"example":"user_inputs/paper-reading-workbench/sources.md"}],"outputs":[{"id":"reading_index","label":"阅读队列与来源索引","path":"literature/skill_reading_workbench/reading_index.md","description":"每篇论文的解析状态、访问等级、阅读优先级、问题和链接。"},{"id":"paper_cards","label":"阅读卡结构记录","path":"literature/skill_reading_workbench/paper_cards.json","description":"论文元数据、卡片字段、页码/section 覆盖、证据等级与未回答问题。"},{"id":"question_answers","label":"定向问题回答","path":"literature/skill_reading_workbench/question_answers.md","description":"面向用户问题的论文级回答、来源锚点、未知项和需要补读的 section。"},{"id":"learning_summary","label":"跨论文学习总结","path":"literature/skill_reading_workbench/cross_paper_summary.md","description":"共同机制、分歧、边界和可进一步比较/综合的线索。"},{"id":"workflow_manifest","label":"阅读工作台清单","path":"literature/skill_reading_workbench/manifest.json","description":"输入解析、PDF 可读性、阅读覆盖、产物和恢复点。"}]}
workflow	{"kind":"integrated","summary":"将不完整的论文输入先解析成可读队列，再按用户问题生成受证据约束的卡片和跨论文总结。","phases":[{"id":"source_contract","label":"论文清单与问题解析","objective":"确认待读对象、阅读问题、优先级和期待深度。","operations":"[Truncated]"},{"id":"access_and_intake","label":"标识解析、PDF 获取与补料","objective":"尝试解析稳定标识和读取 PDF；缺失时请用户上传或修正，而非用模型记忆替代。","operations":"[Truncated]","human_gate":true},{"id":"evidence_reading","label":"分层阅读与问题取证","objective":"完整/部分/摘要/metadata 分级阅读，记录页码和 section 覆盖。","operations":"[Truncated]"},{"id":"learning_summary","label":"跨论文学习总结","objective":"汇总共同机制、差异、边界和未解决问题，不把摘要当作全文证据。","operations":"[Truncated]"}]}

Paper Reading Workbench

Prioritize papers according to the researcher’s stated question, not only citation or search ranking. For each paper state access level, page/section coverage, extraction status, evidence level, and open uncertainty. A missing PDF is not permission to infer methods or results from metadata. Abstract-only readings may produce cautious coverage hints but cannot answer detailed mechanism, dataset, metric, or causal questions.