fix: schedule weekly advisory after the closed UTC cutoff

.github/workflows/publish_advisory_site.yml

@@ -48,7 +48,7 @@ on:
         default: ""
         type: string
   schedule:
-    - cron: "0 13 * * 6"
+    - cron: "0 13 * * 0"
 
 permissions:
   contents: read

.github/workflows/weekly_advisory_review.yml

@@ -43,7 +43,7 @@ on:
         default: ""
         type: string
   schedule:
-    - cron: "30 12 * * 6"
+    - cron: "30 12 * * 0"
 
 permissions:
   contents: read

docs/data_factor_roadmap.md

@@ -80,6 +80,39 @@ Priority order:
 6. Macro/risk regime: VIX, rates, dollar, credit spreads, oil, yield curve, and
    sector beta.
 
+## 2026-10-06 input publication recovery (planned)
+
+Read-only evidence shows that the Advisor still builds and publishes, but its
+checked-out upstream inputs do not reflect the latest generated artifacts.
+The [October 6 report build](https://github.com/QuantStrategyLab/QuantAdvisorResearch/actions/runs/37460712408)
+has `as_of=2026-10-03`, 42 market-confirmation rows, and zero final picks. The
+rejections are `ai_signal_provenance_untrusted` and
+`theme_momentum_stale_as_of`; these gates must remain unchanged.
+
+- [x] Separate generation, protected publication, and consumer acceptance:
+  RSCP's [fresh theme artifact](https://github.com/QuantStrategyLab/ResearchSignalContextPipelines/actions/runs/37132550227)
+  was retained after a protected-main push rejection; PETR produced a manual
+  publication handoff; the latest AI request failed without pushing files
+- [x] Prepare and locally test the two weekly Advisor schedules for Sunday
+  12:30/13:00 UTC, retaining their clock times, closed-period helper, and freshness
+  contract; remote publication remains pending
+- [x] Record the RSCP generation-versus-publication distinction and fixed-artifact
+  review requirements; workflow identity changes remain deferred
+- [ ] Review the exact existing theme artifact and publish it through a normal
+  PR only after its producer metadata, bytes, time bounds, and policy pass
+  verification; verify the actual PR head's required `test` check
+- [ ] Resolve event-entity evidence and AI signal/manifest recovery in their
+  owning workstreams, without broadening this bounded change
+- [ ] After separately authorized publication, verify the upstream commit or
+  artifact digest actually consumed, report freshness, and deployed report hash
+
+A successful generation run is not completed publication. A one-time reviewed
+PR restores one snapshot only; it does not establish unattended weekly delivery.
+Zero final picks remain an acceptable result when valid evidence is insufficient.
+See [the recovery constraints](system_design.md#input-publication-recovery-2026-10-06)
+for the verified identity and timing limits. This checklist records planned work,
+not a completed repair.
+
 ## Low-Risk Implementation Order
 
 1. Keep public output focused on final recommendations. Preserve

docs/data_factor_roadmap.zh-CN.md

@@ -1,5 +1,23 @@
 # 数据源与因子完善路线
 
+## 2026-10-06 输入发布恢复（待实施）
+
+只读核查确认投顾仍在生成和发布，但 checkout 的上游 main 文件没有跟上新产物。
+[10/06 构建](https://github.com/QuantStrategyLab/QuantAdvisorResearch/actions/runs/37460712408)
+报告截止日为 10/03，有 42 条市场确认，最终推荐和观察均为 0；质量告警为
+`ai_signal_provenance_untrusted` 和 `theme_momentum_stale_as_of`，保留这些门槛。
+
+- [x] 区分生成、受保护发布和消费者接受：RSCP 新主题因保护规则拒绝推 main 后只留 artifact；PETR 只产人工交接；最新 AI 请求失败且未推文件
+- [x] 已准备并本地测试两条周任务的 UTC 周日 12:30/13:00 候选，保留原有时分、周期算法和时间合同；远端发布待处理
+- [x] 已记录 RSCP 生成与发布的区别及固定 artifact 审核要求；持续发布身份改动暂缓
+- [ ] 对固定主题 artifact 验证生产者元数据、原始字节、时间和政策后，另行审核正常 PR；核实该 PR 实际 head 的必需 test
+- [ ] 事件实体证据与 AI signal/manifest 恢复由各自负责工作线处理，不扩大本次有限写集
+- [ ] 发布另获授权后，核验真正消费的上游 commit 或 artifact digest、报告新鲜度及部署报告 hash
+
+一次正常 PR 只恢复一个快照，不代表每周自动交付闭环。有效证据不足时，最终数量为 0
+仍是合法结果；不得降低门槛或改写历史生成时间来制造非零结果。身份和截止时点约束见
+[系统设计恢复说明](system_design.zh-CN.md#2026-10-06-输入发布恢复)。当前清单表示计划，并非修复已完成。
+
 [English](data_factor_roadmap.md) | [简体中文](data_factor_roadmap.zh-CN.md)
 
 ## 当前结论

docs/system_design.md

@@ -168,6 +168,51 @@ monthly, and publish workflows generate it automatically. Short-horizon gates
 require market confirmation, medium-horizon gates are led by theme and symbol
 momentum, and long-horizon gates require durable AI shadow or context strength.
 
+## Input Publication Recovery (2026-10-06)
+
+The bounded recovery changes scheduling and documents publication handoff. It does
+not change renderers, score thresholds, account boundaries, or source contracts.
+The [recovery checklist](data_factor_roadmap.md#2026-10-06-input-publication-recovery-planned)
+tracks implementation and acceptance separately.
+
+At incident capture, the weekly jobs ran on Saturday while `default_weekly_as_of()` selects
+the last *fully closed* Saturday. For example, a 2026-10-03 Saturday run selects
+2026-09-26, whose `reference_time` is 2026-09-27 00:00 UTC. A theme generated on
+2026-10-03 must not pass that earlier cutoff. The local candidate schedules weekly
+review at 12:30 UTC Sunday and publication at 13:00 UTC Sunday, preserving the
+existing times. Both use
+the same helper after 00:00 UTC, so the new report's cutoff includes Saturday's
+already-generated context. A delayed run must still use authentic source dates;
+changing `generated_at`, expiry, or freshness limits is not a recovery mechanism.
+
+`ResearchSignalContextPipelines` is public, and current workflows check out its
+public Git content. That is not proof that the Advisor's job token can download
+cross-repository Actions artifacts. The documented artifact download API needs
+Actions read access to the producer repository. The current workflow exposes no
+verified publisher identity or token with that target-repository permission.
+Connector read access during investigation is not a substitute for CI access.
+No cross-repository artifact reader is enabled by this plan.
+
+The current theme job has `contents: write` only. It cannot be assumed to create
+PRs, and required checks cannot be assumed to run automatically. GitHub's current
+[`GITHUB_TOKEN` documentation](https://docs.github.com/en/actions/concepts/security/github_token)
+says PR opened/synchronize/reopened events created with that token produce
+approval-required runs. A normal reviewed PR must still have its actual head's
+required checks completed. Do not change branch protection or grant new access
+as part of this recovery.
+
+An exact retained artifact may be considered for a separately reviewed one-time
+PR. Before import, verify repository ID, approved workflow path and source SHA,
+same-repository `main` run, event, conclusion, run attempt, artifact ID/name,
+expiry, archive digest, exact member path and file hash, and the snapshot's
+schema/policy/coverage/time contract. Extraction alone proves none of these.
+Record generated, review-required, published, and consumer-accepted as distinct
+states. A reviewed PR does not solve the remaining unattended CI identity gap.
+
+References: [artifact API](https://docs.github.com/en/rest/actions/artifacts#download-an-artifact),
+[fixed producer run](https://github.com/QuantStrategyLab/ResearchSignalContextPipelines/actions/runs/37132550227),
+[consumer build](https://github.com/QuantStrategyLab/QuantAdvisorResearch/actions/runs/37460712408).
+
 ## Fixture vs Live Inputs
 
 Reports built from `examples/` are `source_mode=fixture` and are suitable for

docs/system_design.zh-CN.md

@@ -2,6 +2,35 @@
 
 [English](system_design.md) | [简体中文](system_design.zh-CN.md)
 
+## 2026-10-06 输入发布恢复
+
+本次计划只修改周任务截止时点并记录主题发布交接，不改网页、评分门槛、账户边界或可信输入合同。
+进度见[恢复清单](data_factor_roadmap.zh-CN.md#2026-10-06-输入发布恢复待实施)。
+
+事件核查时，周六运行却选“最近完全闭合的周六”：10/03 的运行选 09/26，参考截止时间为
+09/27 00:00 UTC，因此 10/03 新生成的主题不能合法用于这一旧截止日。本地候选将两条周任务
+移到 UTC 周日 12:30/13:00，保留原有时分和周期函数，使当周六生成的上下文处于报告截止前。
+延迟执行仍按真实来源时间验证，不补假日期、不延长 freshness 或 expiry 来消除告警。
+
+RSCP 是公开仓库；当前 QAR 能 checkout 公开 Git 内容，不代表其 job token 能读取
+RSCP 的 Actions artifact 下载。下载 API 要求生产者仓库的 Actions read 权限，当前
+工作流没有已验证的跨仓库读取身份。诊断连接器读取成功不能替代 CI 身份证明，本方案
+暂不新增消费者 artifact 下载器。
+
+主题任务当前只有 contents:write，不能假定可开 PR。按 GitHub 当前官方文档，
+GITHUB_TOKEN 创建或更新 PR 会产生 approval-required 的 opened/synchronize/reopened
+运行，仍须人工批准并核验实际 PR head 的必需检查；不以修改分支保护或扩大权限解决。
+
+固定既有 artifact 的一次正常 PR 导入须另行审核：核实仓库 ID、批准的 workflow 路径与
+源码 SHA、同仓 main、event、run attempt、成功结论、artifact ID/name/有效期、压缩包
+digest、唯一文件路径和 hash，以及快照 schema/policy/覆盖率/时间合同。解压复制不等于
+可信通过。生成、待审、发布、被消费者接受是不同状态；一次 PR 不代表持续自动化已恢复。
+
+依据：[固定生产运行](https://github.com/QuantStrategyLab/ResearchSignalContextPipelines/actions/runs/37132550227)、
+[GITHUB_TOKEN 触发规则](https://docs.github.com/en/actions/concepts/security/github_token)、
+[artifact 下载规则](https://docs.github.com/en/rest/actions/artifacts#download-an-artifact)。
+
+
 ## 当前架构理解
 
 QuantStrategyLab 现有仓库已经天然分层：

tests/test_build_pipeline.py

@@ -15,6 +15,7 @@
     default_weekly_as_of,
 )
 from quant_advisor_research.cross_repo_smoke import run_cross_repo_smoke
+from quant_advisor_research.time_contract import assess_context_freshness, report_time_bounds
 
 
 ROOT = Path(__file__).resolve().parents[1]
@@ -57,6 +58,56 @@ def test_default_as_of_values_use_only_closed_periods() -> None:
     assert default_monthly_as_of(dt.date(2026, 6, 20)) == dt.date(2026, 5, 31)
 
 
+@pytest.mark.parametrize(
+    ("run_date", "expected_as_of"),
+    [
+        (dt.date(2026, 10, 3), dt.date(2026, 9, 26)),
+        (dt.date(2026, 10, 4), dt.date(2026, 10, 3)),
+        (dt.date(2026, 1, 3), dt.date(2025, 12, 27)),
+        (dt.date(2026, 1, 4), dt.date(2026, 1, 3)),
+        (dt.date(2026, 2, 1), dt.date(2026, 1, 31)),
+        (dt.date(2026, 11, 1), dt.date(2026, 10, 31)),
+    ],
+)
+def test_weekly_cutoff_keeps_saturday_closed_and_rolls_on_sunday(
+    run_date: dt.date, expected_as_of: dt.date
+) -> None:
+    assert default_weekly_as_of(run_date) == expected_as_of
+    generated_at = dt.datetime.combine(run_date, dt.time(12, 30), tzinfo=dt.UTC)
+    bounds = report_time_bounds(expected_as_of, generated_at)
+    assert bounds.reference_time == dt.datetime.combine(
+        expected_as_of + dt.timedelta(days=1), dt.time.min, tzinfo=dt.UTC
+    )
+    assert bounds.reference_time <= generated_at
+
+
+def test_sunday_weekly_cutoff_accepts_saturday_context_without_rewriting_its_time() -> None:
+    context = {
+        "as_of": "2026-10-02",
+        "generated_at": "2026-10-03T11:45:00Z",
+        "expires_at": "2026-10-09T23:59:59Z",
+    }
+    for run_date, expected_reason in (
+        (dt.date(2026, 10, 3), "as_of_in_future"),
+        (dt.date(2026, 10, 4), "fresh"),
+    ):
+        as_of = default_weekly_as_of(run_date)
+        bounds = report_time_bounds(
+            as_of, dt.datetime.combine(run_date, dt.time(13), tzinfo=dt.UTC)
+        )
+        freshness = assess_context_freshness(
+            context,
+            report_as_of=as_of,
+            reference_time=bounds.reference_time,
+            report_generated_at=bounds.generated_at,
+            max_age_days=7,
+        )
+        assert freshness.reason == expected_reason
+        assert freshness.valid is (expected_reason == "fresh")
+
+    assert context["generated_at"] == "2026-10-03T11:45:00Z"
+
+
 def test_build_advisory_artifacts_builds_market_report_and_site(tmp_path: Path) -> None:
     result = build_advisory_artifacts(
         as_of=dt.date(2026, 5, 31),

tests/test_workflow_defaults.py

@@ -1,11 +1,29 @@
 from __future__ import annotations
 
 from pathlib import Path
+import re
+
+import pytest
 
 
 ROOT = Path(__file__).resolve().parents[1]
 
 
+@pytest.mark.parametrize(
+    ("workflow", "expected_cron"),
+    [
+        ("weekly_advisory_review.yml", "30 12 * * 0"),
+        ("publish_advisory_site.yml", "0 13 * * 0"),
+    ],
+)
+def test_weekly_workflows_run_after_saturday_closes_without_changing_utc_time(
+    workflow: str, expected_cron: str
+) -> None:
+    text = (ROOT / ".github" / "workflows" / workflow).read_text(encoding="utf-8")
+    assert re.findall(r'^\s+- cron: "([^"]+)"$', text, flags=re.MULTILINE) == [expected_cron]
+    assert "default_weekly_as_of" in text
+
+
 def test_public_workflows_default_to_live_event_inputs() -> None:
     for workflow in ("weekly_advisory_review.yml", "publish_advisory_site.yml", "monthly_advisory_review.yml"):
         text = (ROOT / ".github" / "workflows" / workflow).read_text(encoding="utf-8")