Current projects,
one public index.

论文风格的双语 HTML 交付，含完整草稿、表格、图表与参考文献。

交互式可视化：银河系参考模型如何投影到 M31 视线并转换为图 3 表面亮度数据。

已接受环带的响应式技术 QA，以及未来径向计划的本地数据覆盖边界。

OBSID 0402560101 的直接 FK5 原生椭圆 RA/Dec QA；视觉批准待定。

已撤回：E021 探测器裁剪响应几何页面；保留作溯源。

作为规范的 XMM/ESAS 谱抽取教学材料，逐步 ESAS 抽取审计。

中文视觉 QA 报告与当前八扇区抽取计划/状态页面。

R0–15 径向计划下的跨相机 EPIC 对比。

PN0 Fe-L 风轴审计；中英双语交付。

M104 温度剖面反馈证据审计。

交互式 idea 进化追踪：周期进度、逐 idea 路径时间线、红队攻击、架构状态。

项目概览：以档案证据、对抗评审、可追溯可行性和 TAC 式判断推动 proposal idea 持续演化。

Elo 排名最高的 proposal idea：NGC 253 SE halo 偏置指向观测，以压力比检验星系风对称性；包含可证伪主张、门控状态与主要风险。

Reveal.js 组会幻灯片：介绍证据驱动的 proposal idea 演化架构、独立 red-team、生命周期门控与双执行 lane。

调度架构、批次时序、实测利用率与并行轴重新设计的静态技术报告。

PI 审阅页：以同半径且旋臂相位匹配的对照，检验 M51–NGC 5195 相互作用轴上的盘面软 X 射线是否存在超出旋臂结构预期的余量。

PI 审阅页：用新 SE 偏置 XMM 指向与档案 NW 指向的点反射匹配孔径，检验 NGC 253 热晕双侧性；明确记录几何、恢复和系统误差开放门槛。

PI 审阅页：在同一半径与精确相同螺旋相位下，比较 M51 B2 与反足 C2_anti 的 diffuse-gas EM；无新 XMM 时间申请，E1/E3/E8/E9/E11 仍开放。

PI 审阅页：检验 M101 D25 是否保留超出方位稳健径向模型的投影弥散 X 射线残差；现只请求 R6 HEAVY 资格化，不申请 XMM 时间，I226-E1–E4 仍开放。

PI 审阅页：用点反射匹配的 SE XMM EPIC 偏移指向，在几何相同孔径内直接测量 NGC 253 热晕 SE/NW 侧向性；检验 Bauer 2008 的不对称是否为孔径假象。

PI 审阅页：现有 M33 XMM EPIC-pn mosaic 是否值得完成经资格化的内外盘投影有效温度对比；不请求新观测时间，response/recovery 与可识别性仍开放。

PI 审阅页：M101 的 source-controlled projected diffuse X-ray 与 local FUV coupling 是否因 spiral-arm status 改变；不请求新观测时间，所有 geometry、source-control、response 与 recovery qualifications 仍开放。

PI 审阅页：NGC 5866 首次 XMM 观测——在冻结的 5 kpc 短轴扇区中检验热 ISM 是否仍延伸；请求 R6 HEAVY 资格化，不申请观测时间，I113-E2/E3 仍开放。

PI 审阅页：五个边缘朝向螺旋星系在共同 116–124 kpc 短轴扇区的表面亮度与恒星质量关联——精确 FK5 多边形、星系特定角变换、探测器包含、源掩膜、有效立体角与响应加权为不可变的 I086-E1 产物，不做全环外推；I086-E1–E5 仍开放。

PI 审阅页：对 NGC 2683（10 Mpc 侧向 Sab）现有 XMM 0671430201（PN 25.029 ks）与 Chandra 11311（39.040 ks）做 source-masked 弥散软 X 射线再提取——检出热 halo、盘主导 ISM，或有约束力的上限；零新观测请求，I108-E1/E2/E3 仍开放。

PI 审阅页：在冻结的 I146-E1 v2 几何契约下，以同单位 δ_D25 与 source-plus-background likelihood，对比 NGC 2146 vs NGC 5907 与 NGC 4594 vs NGC 4666 的正短轴 ±15° 扇区 80–120 kpc 投影表面亮度；仅请求 R6 HEAVY 资格化，不申请 XMM 时间，I146-E1/E2/E3/E4 仍开放。

PI 审阅页：检验 Richings+2010 的 Chandra 热 ISM 结果能否转移到现行 source-masked XMM EPIC per-camera 算子——约束 pair-scale 投影软成分 A_pair 或给出合格上界；零新观测请求，I173-E2/E3 仍开放。

PI 审阅页：在冻结的 PA=[-30°,+30°] 扇区与 5–20′ 径向分箱内，检验 IC 10 光学体之外的软 X 射线投影径向剖面能否相对于半径变化的绝对背景模型被测量；仅请求 R6 HEAVY 资格化，不申请 XMM 时间，I176-E3/E4/E5/G1 仍开放。

中英双语并排 v30.3 / v30.4 / v30.5 正文，加逐版本 What's New 说明。

M82 多相星暴风物理页面；双语交付。

三分区 M82 模拟谱与可执行教程。

M82 三分区模拟谱的 Chandra 叠加；可执行教程。

M82 星暴风 50 ks 模拟谱教程；双语交付。

HEAGOF σ(ΔC) 预测的分步教程；双语交付。

HEAGOF 校正 Z 检验用于 cstat 拟合优度评估的双语教程。

NGC 3079 的冻结 P1/P2 严格 TDD 基准报告。

算法 tournament、真实 EPIC 数据接入、响应折叠与 P4 科学测量状态的进展报告。

11 节方法论课程，涵盖信息预算、过拟合识别、样本外检验、基准构建、指标设计、组合分散化、回测对账与 PBO/DSR，全部使用真实项目数据。

中英双语个人入口，连接科研 Pages 总目录、XMM Proposal Agent Dashboard 与 Handmux 远程终端。

huangrui.pages.dev 的字节级相同镜像，提供同一中英双语个人入口。

xsherpa 拟合排错与开发讨论的交互式回放，支持逐块播放、Thinking/Tools 筛选和文件导航。

Hermes Agent 工作会话的交互式回放，展示用户消息、assistant 响应、tool calls 与文件引用。

包含中英双语网页、llms.txt、agent execution contract、可校验下载资产，以及保留原 statusline 的可回滚安装器。

文献 evidence-delta ledger：严格区分当日新论文与历史回看，并显示 provisional question 路由。

对同一 full-sample display subset 的理论与 catalog 经验条件均值和 residual width 进行比较，并明确 selection 与覆盖边界。

双语冻结快照，报告近周系统进展、portfolio 路由、执行阻塞、草稿边界与下一步资格闭环。

NGC 253 匹配 7–12′ SE/NW 热压比测量的提案展示页，含几何验证、文献边界与系统误差分析。

17 个径向 region 的可审查 XMM MOS 光谱总览：每张谱均来自保存的独立 SrcApec 最佳拟合，并保留逐数据集的 step best fit、残差与平滑总模型。

初步 X-ray → 热 SZ y 推断说明：明确列出输入 annuli、假设、系统边界与未包含项。

CDFS 重复观测的软质子背景拟合审计：对比旧单幂律与 XARTATOMS BPL 约定，含逐谱联合拟合、成分分解与残差诊断。

用当前 SAS ≥20 + ESAS ≥3 标定重做 Strickland 2007 的 NGC 6810 分析，检验盘面外软发射、径向表面亮度轮廓与温度/丰度是否可复现；三种结局（SUPPORT/WEAKEN/INCONCLUSIVE）均明确定义。

按 region 浏览全部可用 ObsID 的静态谱页，含联合拟合结果（a=40–48′ MWhalo+MWhotHalo）、径向 XSherpa inspect 对照与无模型径向 band-flux profile。

检验矮星暴星系热气体是否延伸至 8–15 kpc halo 的 PI 审阅页；含科学动机、低质量端 L_X,halo–M_star 标定意义与当前状态。

用当前 SAS/ESAS 定标重检验 Strickland 2007 的 NGC 6810 盘外软 X 射线结果；历史正对照设计，三种结局（SUPPORT/WEAKEN/INCONCLUSIVE）均明确定义。

用当前 SAS ≥20 + ESAS ≥3 校准重处理 NGC 6810，检验盘面外 ≥5 kpc 软发射的 3σ 可检测性与温度/丰度约束；三种可证伪结局均明确定义。

检验 NGC 1808 短轴 18–30′ 扇区投影软 X 射线对比的 PI 审阅页；含科学动机、天空几何、测量定义、四种可证伪结局与成本边界。

I145 的姊妹 PI 审阅页：聚焦外侧 18–22′ vs 26–30′ 扇区对比，强调任何单调径向剖面已产生正对比的判别陷阱，以及 NGC 1808–NGC 1827 pair environment 的投影混淆。

检验高度侧向星暴星系 NGC 4700 的 minor-axis 绝对软 X 射线成分存在性；含背景模型冻结、四种可证伪结局、温-热对应关系意义与 R6 HEAVY 资格化请求。

检验 Virgo 坠入旋涡星系 NGC 4634 两侧 minor-axis 热 halo 的绝对存在性；含两世界简并证明、环境 CGM 剥离模型依赖、四种互斥判定（PRESENT/BOUNDED ABSENT/PRESENT-BUT-UNATTRIBUTED/INCONCLUSIVE）与 R6 HEAVY 资格化请求。

7 张可交互架构图：核心架构、idea 生命周期、HEAVY 授权时序、Cognition OS 数据流、idea 状态机、pipeline 角色与 cycle 演化。

逐步拆解 2.1 filter 的 Al/Si 线扫描策略，揭示 sample construction 导致的表观不连续性；含 200 个 trial 的 CSV 与诊断 JSON 下载。

静态证据审计：122 个 Q 文件未构成问题图（0 条 typed Q→Q 边），P DAG 仅 2 条声明弧，Q→P 交叉引用仅 7 条。

0.7 keV 辐射起源的可审查空间分解草稿：一句话论断、谱模型契约、三种候选起源与尚未解决的部分。

6 张可交互架构图：核心架构、阅读循环、分歧生命周期、Ledger 数据流、引擎路由与配额、诚实守则。

Manuscript-style bilingual HTML delivery with the complete draft, tables, figures, and references.

Interactive visualization of how the Milky Way reference model is projected onto the M31 line of sight and converted to Figure 3 surface-brightness data.

Responsive technical QA for accepted annuli plus a local-data coverage boundary for the future radial program.

Direct FK5 native-ellipse RA/Dec QA for OBSID 0402560101; visual approval pending.

Withdrawn: E021 detector-clipped response geometry page; retained for provenance.

Step-by-step ESAS extraction audit used as the canonical XMM/ESAS spectrum extraction teaching material.

Chinese visual QA report and the current eight-sector extraction-plan/status page.

Cross-camera EPIC comparison across the R0–15 radial program.

PN0 Fe-L wind-axis audit; Chinese/English bilingual delivery.

M104 temperature-profile feedback evidence audit.

Interactive idea-evolution trace: cycle progress, per-idea journey timeline, red-team shots, and architecture status.

Project overview of an evidence-driven, adversarially reviewed framework for continuously evolving proposal ideas with traceable feasibility and TAC-style judgment.

Top Elo-ranked proposal idea: NGC 253 SE halo offset pointing to test wind symmetry via pressure ratio; includes falsifiable claim, gate status, and key risks.

Reveal.js group-meeting deck on evidence-driven proposal-idea evolution, independent red-teaming, lifecycle gates, and dual execution lanes.

Static technical report on scheduler architecture, batch timing, measured utilization, and redesigning the axis of parallelism.

PI-review page for an M51 tidal-axis contrast measured against same-radius, arm-phase-matched disk controls.

PI-review page for a point-reflected SE/NW XMM comparison of NGC 253's hot halo, with geometry, recovery, and systematic gates explicitly left open.

PI-review page for a same-radius, exactly phase-matched diffuse-gas EM contrast between M51 B2 and antipodal C2_anti; no new XMM time is requested and E1/E3/E8/E9/E11 remain open.

PI-review page asking whether M101's D25 boundary retains a projected diffuse-X-ray residual beyond an azimuthal-robust radial model; it requests only R6 HEAVY qualification, not XMM time, while I226-E1–E4 remain open.

PI-review page: a point-reflected SE XMM EPIC offset pointing to measure NGC 253 hot-halo SE/NW sidedness in geometrically identical apertures; tests whether the Bauer 2008 asymmetry survives geometric control.

PI-review page asking whether the M33 XMM EPIC-pn mosaic warrants qualification of a projected inner–outer effective-temperature contrast; no new XMM time is requested and response/recovery/identifiability remain open.

PI-review page asking whether M101's source-controlled projected diffuse-X-ray/local-FUV coupling changes with spiral-arm status; no new XMM time is requested and all geometry, source-control, response, and recovery qualifications remain open.

PI-review page for NGC 5866's first XMM observation — testing whether hot ISM remains extended at a frozen 5 kpc minor-axis sector; requests R6 HEAVY qualification, not observing time, while I113-E2/E3 remain open.

PI-review page for a five edge-on spiral-galaxy sample at a common 116–124 kpc minor-axis sector, asking whether the censored S_X–M_star correlation is testable; precise FK5 polygons, galaxy-specific angular transforms, detector inclusion, source masking, effective solid angle, and response weighting are produced as immutable I086-E1 deliverables, with no full-annulus extrapolation, while I086-E1–E5 remain open.

PI-review page for a source-masked re-extraction of diffuse soft X-ray emission from the existing XMM 0671430201 (PN 25.029 ks) and Chandra 11311 (39.040 ks) observations of NGC 2683 (edge-on Sab at 10 Mpc) — a hot-halo detection, a disk-dominated ISM detection, or a constraining upper limit; no new observing time is requested and I108-E1/E2/E3 remain open.

PI-review page for two pre-declared contrasts (NGC 2146−NGC 5907, NGC 4594−NGC 4666) of the positive-minor-axis ±15° sector at 80–120 kpc projection, under the frozen I146-E1 v2 geometry contract with per-camera δ_D25 and source-plus-background likelihood; requests only R6 HEAVY qualification, no XMM time, while I146-E1/E2/E3/E4 remain open.

PI-review page asking whether the Richings+2010 Chandra hot-ISM result transfers to a current source-masked XMM EPIC per-camera operator — constraining the pair-scale projected soft component A_pair or placing a qualified bound; no new observing time is requested and I173-E2/E3 remain open.

PI-review page asking whether IC 10's soft X-ray projected radial profile beyond the optical body can be measured against a radius-varying absolute background model within a frozen PA=[-30°,+30°] sector and 5–20′ radial bins; requests only R6 HEAVY qualification, no XMM time, while I176-E3/E4/E5/G1 remain open.

Side-by-side v30.3 / v30.4 / v30.5 main text in Chinese and English, plus per-version What's New notes.

Multiphase M82 starburst-wind physics page; bilingual delivery.

Three-region simulated M82 spectrum with executable tutorial.

M82 three-region simulated spectrum with Chandra overlay; executable tutorial.

M82 starburst-wind 50 ks mock-spectrum tutorial; bilingual delivery.

Step-by-step tutorial on HEAGOF σ(ΔC) prediction; bilingual delivery.

Bilingual tutorial on HEAGOF corrected Z-test for cstat goodness-of-fit evaluation.

Frozen P1/P2 strict-TDD benchmark report for NGC 3079.

Progress report covering the algorithm tournament, real EPIC data ingestion, response folding, and P4 science-measurement status.

11 methodology lessons covering information budget, overfitting detection, out-of-sample testing, benchmark construction, metric design, portfolio diversification, backtest reconciliation, and PBO/DSR, all using real project data.

Bilingual personal entry linking the research Pages directory, the XMM Proposal Agent Dashboard, and the Handmux remote terminal.

Byte-identical mirror of huangrui.pages.dev, providing the same bilingual personal entry.

Interactive replay of an xsherpa fitting-debug and development discussion, with block playback, Thinking/Tools filters, and file navigation.

Interactive replay of a Hermes Agent work session, showing user messages, assistant responses, tool calls, and file references.

Bilingual pages, llms.txt, an agent execution contract, verifiable download assets, and a reversible installer that preserves the existing statusline.

Evidence-delta ledger separating daily new papers from historical review and showing provisional-question routing.

Comparison of theoretical and catalog-empirical conditional means and residual widths on the same full-sample display subset, with explicit selection and coverage limits.

Bilingual frozen snapshot reporting recent system progress, portfolio routing, execution blockers, draft boundaries, and the next qualification steps.

Presenter page for the matched 7–12′ SE/NW thermal-pressure ratio measurement in NGC 253, with geometry verification, literature bounds, and systematic-floor analysis.

Auditable overview of XMM MOS spectra across 17 radial regions; every plot derives from a saved, independent SrcApec best fit with per-dataset step best fits, residuals, and the smoothed total model.

Preliminary X-ray → thermal-SZ y inference note: explicitly lists input annuli, assumptions, systematic boundaries, and what is deliberately excluded.

Soft-proton background fit audit for repeated CDFS observations: compares the deprecated single-power-law trial against the XARTATOMS BPL convention, with per-spectrum joint fits, component decomposition, and residual diagnostics.

Reprocessing Strickland 2007's NGC 6810 analysis with current SAS ≥20 + ESAS ≥3 calibration to test off-plane soft emission, radial surface-brightness profile, and temperature/abundance reproducibility; all three outcomes (SUPPORT/WEAKEN/INCONCLUSIVE) are explicitly defined.

Per-region browsing of all available ObsID static spectrum pages, including joint-fit results (a=40–48′ MWhalo+MWhotHalo), radial XSherpa inspect comparison, and model-free radial band-flux profiles.

PI-review page for testing whether hot gas in a dwarf starburst extends to the 8–15 kpc halo; includes science motivation, low-mass-end L_X,halo–M_star calibration significance, and current status.

Re-testing Strickland 2007's NGC 6810 off-plane soft X-ray results with current SAS/ESAS calibration; historical positive-control design with all three outcomes (SUPPORT/WEAKEN/INCONCLUSIVE) explicitly defined.

Reprocessing NGC 6810 with current SAS ≥20 + ESAS ≥3 calibration to test 3σ detectability of ≥5 kpc off-plane soft emission and temperature/abundance constraints; all three falsifiable outcomes explicitly defined.

PI-review page for the projected soft-X-ray contrast in NGC 1808's 18–30′ minor-axis sectors; includes science motivation, sky geometry, measurement definition, four falsifiable outcomes, and cost boundary.

Companion PI-review page for I145: focuses on the exterior 18–22′ vs 26–30′ sector contrast, highlighting the discriminant trap that any monotonic radial profile already yields a positive contrast, plus the NGC 1808–NGC 1827 pair-environment projection ambiguity.

Testing absolute soft X-ray component existence in the minor-axis of edge-on starburst NGC 4700; includes frozen background model, four falsifiable outcomes, warm-hot correspondence significance, and R6 HEAVY qualification request.

Testing absolute existence of hot halo on both minor-axis sides of Virgo-infalling spiral NGC 4634; includes two-world degeneracy proof, environmental CGM stripping model dependencies, four mutually exclusive verdicts (PRESENT/BOUNDED ABSENT/PRESENT-BUT-UNATTRIBUTED/INCONCLUSIVE), and R6 HEAVY qualification request.

7 interactive architecture diagrams: core architecture, idea lifecycle, HEAVY authorization sequence, Cognition OS data flow, idea state machine, pipeline roles, and cycle evolution.

Step-by-step dissection of the 2.1 filter Al/Si line-scan strategy, revealing apparent discontinuities from sample construction; includes CSV of 200 trials and diagnostic JSON downloads.

Static evidence audit: 122 Q files form no question graph (0 typed Q→Q edges), the P DAG has only 2 declared arcs, and the Q→P bridge is 7 literal mentions.

Auditable spatial-decomposition draft of the 0.7 keV emission origin: one-line claim, spectral-model contract, three candidate origins, and unresolved items.

6 interactive architecture diagrams: core architecture, reading loop, disagreement lifecycle, ledger data flow, engine routing and quota, and honesty rules.