目录Contents
当前状态（2026-10-14）Current state (2026-10-14)
每一条都标注可信级别（Trusted / Provisional / Running / Questioned / Deprecated），与 cognition-os/03_TRUST_TABLE.md 一致。
Every item carries its trust level (Trusted / Provisional / Running / Questioned / Deprecated), matching cognition-os/03_TRUST_TABLE.md.
已确立（Trusted）Established (Trusted)
| 内容Item | 数值Number | 来源Source |
|---|---|---|
| v19 基线（14 主场）共温v19 baseline common temperature (14 primary fields) | kT = 0.17531 ± 0.00436 keV；异质性检验 Q(13)=11.631，p=0.558kT = 0.17531 ± 0.00436 keV; heterogeneity test Q(13)=11.631, p=0.558 | cognition-os/insights/I012；03_TRUST_TABLE.md「Production v19 independent CGMsum baseline」cognition-os/insights/I012; 03_TRUST_TABLE.md "Production v19 independent CGMsum baseline" |
| CGMsum 全场吸收流量（0.5–2 keV 逆方差）CGMsum all-field absorbed flux (0.5–2 keV, inverse variance) | 0.965 ± 0.044 | cognition-os/insights/I033 第 1 条cognition-os/insights/I033 item 1 |
| 轮廓平（常数拟合）The profile is flat (constant fit) | χ² = 11.99/13，p = 0.53 | newangels3_discussion/quick_shape_limit_20261014.txt 第 1 行；03_TRUST_TABLE.md「CGMsum profile flat」newangels3_discussion/quick_shape_limit_20261014.txt line 1; 03_TRUST_TABLE.md "CGMsum profile flat" |
| 样本与几何（22 指向、43 探测器组、半径 10.5–30.8 kpc）Sample and geometry (22 pointings, 43 detector sets, radii 10.5–30.8 kpc) | 22 / 43 / 10.5–30.8 kpc | cognition-os/insights/I005；03_TRUST_TABLE.md 第 1 行cognition-os/insights/I005; 03_TRUST_TABLE.md row 1 |
| 「MWhalo」标签 = MW+M31 未分解之和The "MWhalo" label = the unresolved MW+M31 sum | SrcApec.norm = 0（冻结）SrcApec.norm = 0 (frozen) | cognition-os/insights/I006；03_TRUST_TABLE.md「MWhalo = unresolved」cognition-os/insights/I006; 03_TRUST_TABLE.md "MWhalo = unresolved" |
| MOS 1.4–2.0 keV 剔除与逐指向 HI4PI N_HMOS 1.4–2.0 keV exclusion and per-pointing HI4PI N_H | N_H 范围 4.81e20–7.64e20 cm^-2N_H range 4.81e20–7.64e20 cm^-2 | cognition-os/experiments/E003「HI4PI QC」；03_TRUST_TABLE.md「Per-OBSID HI4PI nH」cognition-os/experiments/E003 "HI4PI QC"; 03_TRUST_TABLE.md "Per-OBSID HI4PI nH" |
暂定（Provisional）Provisional
| 内容Item | 数值Number | 边界Boundary |
|---|---|---|
| 集中式 M31 β 晕上限（r_c=4 kpc，R=20 kpc）Upper limit on a concentrated M31 beta halo (r_c=4 kpc, R=20 kpc) | 近似 95% 上限 0.12/0.08/0.05（β=0.4/0.5/0.6，MW 常数；D49 后默认）；MW 线性 b 版 0.19/0.12/0.08 所依赖的纬度梯度已被 E014 排除approximate 95% UL 0.12/0.08/0.05 (β=0.4/0.5/0.6, MW constant; the default after D49); the MW linear-in-b version, 0.19/0.12/0.08, relied on the latitude gradient that E014 excluded | 流量级二阶拟合、局部协方差误差、近似高斯 95%；权威版本是 E013flux-level second-stage fit, local-covariance errors, approximate Gaussian 95%; the authoritative version is E013 |
| 与 Zhang 2024 叠加比较Comparison with the Zhang 2024 stack | Zhang 预期 0.83 @20 kpc，比 MW 常数版上限（0.05–0.12）高 7–17 倍Zhang predicts 0.83 at 20 kpc, 7–17 times above the MW-constant limit (0.05–0.12) | 叠加/模拟均值转移到吸收 0.5–2 keV；模型比较，不是排除stack/simulation means transferred to absorbed 0.5–2 keV; a model comparison, not a rejection |
| 外 NW 四场偏亮（原纬度梯度 Δχ²=6.2）The outer-NW four are bright (the original latitude-gradient Δχ²=6.2) | 外四 1.114 ± 0.071 vs 其余十 0.870 ± 0.056；E014：RASS 比值 0.91 ± 0.12（无大尺度结构）、OMNI 无窗口超 4e8（非 SWCX）、SP 相关性弱（+0.35，p=0.22）the outer four 1.114 ± 0.071 vs the other ten 0.870 ± 0.056; E014: RASS ratio 0.91 ± 0.12 (no large-scale sky), no OMNI window above 4e8 (not SWCX), weak SP correlation (+0.35, p=0.22) | 按 D49 记为未解释场级系统误差（~0.25）；重拟合（SP 谱形自由）与更严格耀斑过滤已排队per D49 carried as an unexplained field-level systematic (~0.25); a refit (SP shape freed) and a stricter flare filter are queued |
| 南北差（0.4–1.25 keV 基线）North-minus-South (0.4–1.25 keV baseline band) | N 1.106 ± 0.054，S 0.894 ± 0.203，差 0.212 ± 0.210（1.01σ）N 1.106 ± 0.054, S 0.894 ± 0.203, difference 0.212 ± 0.210 (1.01σ) | 两侧半径分布不匹配（中位 19.28 vs 14.97 kpc）；描述性平均，非联合拟合unmatched side radius distributions (medians 19.28 vs 14.97 kpc); descriptive means, not a joint fit |
| Locatelli 模型紧张The Locatelli model tension | 模型 1.290 vs 测量 0.965；残差 −0.326model 1.290 vs measured 0.965; residual −0.326 | 命名模型的相容性诊断，不是负的物理发射；三种解释并列（模型高估/转换失配/总和偏低）a named-model incompatibility diagnostic, not negative physical emission; three live explanations (model over-prediction / conversion mismatch / low-biased sum) |
来源：cognition-os/insights/I033；newangels3_discussion/quick_shape_limit_20261014.txt 第 2–6 行与 β 表；cognition-os/insights/I012；外 NW 行另据 cognition-os/experiments/E014（2026-10-14T07:40Z 结果节）与 cognition-os/04_DECISION_LOG.md D49；03_TRUST_TABLE.md 对应行。
Source: cognition-os/insights/I033; newangels3_discussion/quick_shape_limit_20261014.txt lines 2–6 and the beta table; cognition-os/insights/I012; the outer-NW row also from cognition-os/experiments/E014 (the 2026-10-14T07:40Z result section) and cognition-os/04_DECISION_LOG.md D49; the matching rows of 03_TRUST_TABLE.md.
正在运行（结果待定）Running (results Pending)
| 实验Experiment | 问题Question | 状态Status |
|---|---|---|
| E013 | 用谱联合似然（而非流量级拟合）重新给出 M31 β 上限；A1 逐场 Δstat 曲面 → A2 空间模型re-derive the M31 beta limit with the spectral joint likelihood (not the flux-level fit); A1 per-field Δstat surfaces → A2 spatial models | A1 running （2026-10-14 启动）(launched 2026-10-14) |
| E014 | 外 NW 超出是天空结构、SWCX 还是仪器？RASS R45 + OMNI + 噪声表is the outer-NW excess sky structure, SWCX, or instrumental? RASS R45 + OMNI + the nuisance table | done 2026-10-14 （结论见上表外 NW 行与 D49；Hermes 校正轮仅为台账补跑）(the result is in the outer-NW row above and D49; the Hermes correction round runs for the ledger only) |
来源：cognition-os/experiments/E013_joint_spatial_profile_likelihood.md；cognition-os/experiments/E014_nw_gradient_rass_omni_crosscheck.md；cognition-os/04_DECISION_LOG.md D49。E013 结果待定——本站不猜测；E014 已完成（结论见上表）。
Source: cognition-os/experiments/E013_joint_spatial_profile_likelihood.md; cognition-os/experiments/E014_nw_gradient_rass_omni_crosscheck.md; cognition-os/04_DECISION_LOG.md D49. E013's result is Pending — this site does not guess it; E014 is done (result in the table above).
被质疑 / 已废弃（摘自信任表）Questioned / deprecated (from the trust table)
| 条目Item | 状态Status | 原因Why |
|---|---|---|
| v18 逐 ObsID 流水线The v18 per-ObsID pipeline | 已废弃Deprecated | B003：有效滤波把已剔除的分组通道带回；仅留作回归参照B003: the effective filter reintroduced excluded grouped channels; kept only as a regression reference |
| 「晕 = 纯银河系天空背景」（D01）"Halo = pure MW sky background" (D01) | 已废弃Deprecated | D26：080073 是 M31 内晕视线，不是空白前景场D26: the 080073 fields are M31 inner-halo sightlines, not blank foreground fields |
| RASS v4/rates 分支与当前 .fitting 会话The RASS v4/rates branch and the current saved .fitting session | 已废弃Deprecated | B001：失败分支覆盖了保存会话，Δstat=+162.27B001: a failed branch overwrote the saved session, Δstat=+162.27 |
| delta_stat F 检验分类（D22）The delta_stat F-test classification (D22) | 被质疑Questioned | 非负边界处 F 检验无效；论文结论需 profile likelihood 或模拟校准an F-test is invalid at the non-negative boundary; paper claims need profile likelihood or simulation calibration |
| Kavanagh 2020 作为晕先例Kavanagh 2020 as halo prior art | 已废弃Deprecated | D48 更正：其区域是盘旁背景，不是晕视线D48 correction: their regions are disc-adjacent backgrounds, not halo sightlines |
| X-LEAP 作为 M31 方向宽带先验X-LEAP as an M31-direction broadband prior | 被质疑Questioned | 无干净行在 M31 4.4° 内；源内部自相矛盾；仅留作验证no clean row within 4.4° of M31; the source is internally inconsistent there; validation only |
来源：cognition-os/03_TRUST_TABLE.md 的对应行（v18 pipeline、Halo as pure MW、RASS v4、delta_stat F-test、Kavanagh、X-LEAP）。
Source: the matching rows of cognition-os/03_TRUST_TABLE.md (v18 pipeline, Halo as pure MW, RASS v4, delta_stat F-test, Kavanagh, X-LEAP).
数据：newangels3_discussion/quick_v19_primary_fields_geometry_20261014.csv（由 independent_cgm_v19_all22_measurements.csv 生成）；常数拟合值来自 quick_shape_limit_20261014.txt。图由 handbook/hermes_figscripts/make_figures.py 生成。
Data: newangels3_discussion/quick_v19_primary_fields_geometry_20261014.csv (generated from independent_cgm_v19_all22_measurements.csv); the constant-fit value from quick_shape_limit_20261014.txt. Figure made by handbook/hermes_figscripts/make_figures.py.