| name | literature-query-plan |
| description | Build a transparent, multilingual and reproducible literature query portfolio from a scoped research brief. Use before web search when the researcher needs explicit source coverage, language policy, inclusion logic, and targeted evidence gaps. |
| tools | ["read_file","write_file","list_files","finish_task"] |
| strict_tools | true |
| model_tier | standard |
| temperature | 0.2 |
| allowed_read_prefixes | ["user_inputs/literature-query-plan/","ideation/","literature/","_runtime/skill_sessions/"] |
| allowed_write_prefixes | ["literature/"] |
| outputs_expected | {"query_plan":"literature/skill_query_plan.md","query_plan_json":"literature/skill_query_plan.json"} |
| interaction | {"mode":"guided","language":"zh-CN","summary":"在实际联网检索前生成可审计的 query 组合、来源/语言策略、去重逻辑和证据缺口计划。","request_required":true,"request_prompt":"请说明要支持的研究问题或论文 section、时间范围、目标语言和希望优先的来源。","example_request":"为英文 Related Work 设计 2022-2026 的 query 组合；正文为英文，保留中文文献作为中国情境或政策证据，不把它们混入英文主检索。","required_inputs":[{"id":"scope","label":"研究范围或检索问题","description":"使用 research-scope 产物或上传独立问题简报；必须写明要回答的证据问题。","paths":"[Truncated]","extensions":"[Truncated]","min_bytes":80,"example":"ideation/skill_scope_brief.md"}],"optional_inputs":[{"id":"current_synthesis","label":"当前综合/缺口","description":"可选；用于排除已覆盖材料并为确切 claim 或 section 定向补检。","paths":"[Truncated]","extensions":"[Truncated]","min_bytes":80,"example":"literature/missing_areas.md"}],"outputs":[{"id":"query_plan","label":"检索计划","path":"literature/skill_query_plan.md","description":"按证据问题、来源、语言、检索式、筛选/去重和停止条件组织的可读计划。"},{"id":"query_plan_json","label":"结构化检索计划","path":"literature/skill_query_plan.json","description":"机器可读的 query portfolio、语言策略和预期覆盖记录。"}]} |
Reproducible Query Planning
Write the declared Markdown and JSON outputs. Define a language policy explicitly: manuscript_language=en means Chinese queries/literature are excluded from the English main-evidence pool unless the request explicitly retains them for a named contextual/policy role; zh mirrors that policy; mixed keeps source language and role per record. Do not claim that a language has been searched yet.
For every query cluster include the evidence question, source(s), query text, language, year range, inclusion/exclusion rules, expected note-card sections, deduplication keys, coverage target, and stop/escalation condition. Reserve broad search for corpus construction; for a single weak claim, prescribe section-level note recovery first.
