Julia_RelaxTime

为 Julia_RelaxTime 的正式数值产物、诊断产物或研究结果生成可追溯分析包。适用于 docs/analysis 分析产物沉淀、输入证据审计、派生表和关键图生成、机制/归因边界判定、claim ledger、论文写作前的 evidence package；可覆盖 transport、relaxtime、phase structure、meson 或其他仓库内数值产物。

在 Julia_RelaxTime 中生产可入库的正式高精度数值产物。适用于 formal production, 正式产物, 高精度产物, 收敛性测试, convergence, production audit, README/manifest, plot_manifest, data/outputs/results, data/outputs/figures。必须先完成收敛性证据，再用证据支持的参数重跑正式产物。

在 Julia_RelaxTime 中执行隔离式文献复现 Spike。适用于指定论文图表、公式或数值结论的独立 tempN 沙箱复刻、文献事实表、未明示口径审计、复现证据链和二选一 verdict：与文献对齐 / 文献信息不足以复现。关键词：isolated reproduction spike, temp, 文献复现, 独立复现, 复刻, figure reproduction, formula audit, verdict

Use when Codex needs to retrieve scholarly literature through INSPIRE-first workflows, including INSPIRE search/sorting, open-access arXiv PDF download, compliant triage between open files and publisher landing pages, and supervised Playwright institution-login handoff for legally authorized full-text access.

为 Julia_RelaxTime 的“提交改动 + 开 PR + 收尾验证”提供标准化流程。适用于用户要求“提交”“开 PR”“收尾”“准备合并”“整理本次改动并发 PR”“处理 review comments 后更新 PR”等场景。该 skill 会先判断是否应优先处理现有 review comments，再按仓库约束完成分支命名、选择性 stage、历史提交风格匹配、分层测试与治理检查、PR 标题与固定正文整理，以及 `docs/dev/active/` 归档检查。

将宽泛的 Julia_RelaxTime 任务收束为可执行 harness：边界、风险、验证、交付物与推荐 skill。适用于跨后端、前端契约、文档、论文、研究分析的混合任务。关键词：task harness, scope, validation, skill routing, execution plan

以 append-only 方式记录 Julia_RelaxTime 的实验、扫描、补点、复测与异常诊断，不重写历史结论。适用于参数扫描、阈值邻域补点、rerun 记录、输出路径留痕。关键词：experiment logbook, append-only, scan, rerun, analysis

为 Julia_RelaxTime 提炼前端可消费的后端数据契约。适用于单点求解、扫描结果、脚本输出、HTTP 接口、JSON/CSV 字段表、单位说明、错误语义、mock payload。关键词：frontend contract, API schema, payload, units, docs/api, docs/dev

为 Julia_RelaxTime 论文写作生成可追溯草稿，只基于源码、结果文件、图表和明确提供的参考材料。适用于方法、实验设置、图注、结果描述、一致性审校。关键词：paper, methods, caption, evidence, traceable, manuscript

管理 Julia_RelaxTime 中 transport / relaxtime 相关改动的回归验证、基线漂移判断与证据汇报。适用于截面、散射率、弛豫时间、输运工作流、固定点基线、smoke/core/full 分层回归。关键词：transport, relaxtime, regression, baseline, drift, tests/regression

为 Julia_RelaxTime 编写或更新 API 文档时使用。适用于按三层视图组织 API 文档：面向用户入口、职责核心、导出 API 全集；并要求导出 API 全集必须通过脚本自动生成而不是人工维护。关键词：API 文档、导出函数、公共接口、职责核心、导出 API 全集、docs/api。

Citation operations skill for DOI/arXiv metadata normalization, BibTeX curation, duplicate cleanup, and reference consistency checks. Use when managing reference quality and bibliography integrity for papers and reports.

Comprehensive cross-domain research synthesis skill focused on broad background gathering, multi-source summaries, and citation-organized overviews. Use for general topic research and perspective mapping, not for implementation-level method selection.

当用户要求按开发文档或任务单连续落地实现时使用。适用于“按文档推进”“按计划执行”“根据任务单落地”“执行文档下一步”“继续推进”“做下一个任务”“按文档自动落地，做完汇报”“不要中间打断，除非卡住或跑偏”“启动 doc-implementation”“使用主线对齐 Skill”。会优先定位 docs/dev/active 下的活动开发文档，按顺序实现、验证、更新状态，并在完成后一次性汇总。

Structured multi-database literature review workflow for academic synthesis with reproducible search logs and thematic evidence mapping. Use when you need rigorous review methodology (scope, screening, synthesis), not general brainstorming.

Structured technical/theoretical research workflow for method selection, model comparison, and implementation-oriented conclusions with validation plans. Use when decisions must map literature evidence to concrete engineering or computational next steps.

OpenAlex-focused scholarly retrieval skill for work/author/institution queries, citation trend snapshots, and topic-level evidence discovery. Use when you need structured literature discovery and bibliometric signals from OpenAlex.

Prefer semantic symbol tools for precise navigation and refactoring. Use vscode_listCodeUsages and vscode_renameSymbol instead of grep-style text search when changing code symbols.

Collaborative technical writing workflow for specs, architecture notes, and research manuscripts. Use when drafting or revising structured documents that require clear audience targeting, iterative refinement, and consistency checks.

Julia 1.10+ engineering guidance for scientific/numerical code, performance optimization, multiple dispatch design, and package-quality implementation. Use when tasks involve Julia code structure, type stability, profiling, testing, or refactoring in computational backends.

Academic and technical writing style guidance focused on clarity, rigor, and human-like narrative flow. Use when drafting papers, polishing manuscripts, rewriting AI-like prose, or improving argument coherence and citation-aware structure.

以 append-only 方式向执行台账追加记录，默认不回读历史正文，降低上下文污染。关键词：执行记录、台账追加、append-only、batch、不回读上下文。

从零设计开发文档（需求文档/技术方案/实施任务单）并输出可勾选结构。适用于新功能立项、重构规划、研究任务分解。关键词：开发文档设计、任务单、DoD、里程碑、验收标准、MVP。

管理数值基线回归的全流程规范（基线生成、存储、测试校验、CI门禁与变更准入），适用于固定点/采样点回归、防止版本更新引入数值漂移。关键词：baseline, regression, fixed points, rtol, atol, smoke, nightly, PNJL, transport

Archive docs/dev/active markdown tasks into docs/dev/archived with required metadata frontmatter, using scripts/dev/archive_docs.jl and validating format compliance.