@geoly-ai/social-hub-cli 0.3.18 → 0.3.20

package/CHANGELOG.md

@@ -4,6 +4,35 @@ All notable changes to `@geoly-ai/social-hub-cli` are documented in this file.
 
 Release workflow: `pnpm release:cli` → `pnpm publish:cli`.
 
+## [0.3.20] - 2026-08-18
+
+### Added
+
+- 发布日历指令端到端接线(Block 5) + 封死 D1 旁路
+- 发布日历指令的 SDK/CLI/skills 交付面(Block 6)
+
+### Changed
+
+- response mock 的 rubricVersion 跟到 content-review.v5
+
+### Documentation
+
+- 写作侧 matrix skill 对齐冻结契约(块 2d)+ 删掉 WeaklyVerified
+
+## [0.3.19] - 2026-08-14
+
+### Added
+
+- classify outcomes instead of reporting 2xx login pages as success
+
+### Fixed
+
+- align AUTHZ_POLICY_VERSION with package version
+
+### Changed
+
+- rebuild reddit matrix SOP around evidence and decision gates
+
 ## [0.3.18] - 2026-08-13
 
 ### Added

package/dist/cmd-manifest.json

@@ -1,6 +1,6 @@
 {
-  "generatedAt": "2026-08-13T06:30:51.783Z",
-  "cliVersion": "0.3.18",
+  "generatedAt": "2026-08-18T09:08:50.530Z",
+  "cliVersion": "0.3.20",
   "commands": [
     "accounts",
     "accounts brand-bindings",
@@ -111,6 +111,8 @@
     "calendar list",
     "calendar list-full",
     "calendar patch",
+    "calendar preflight",
+    "calendar set-delivery-mode",
     "campaigns",
     "campaigns accounts",
     "campaigns accounts assign",

package/dist/content-review.test.js

@@ -73,7 +73,7 @@ const okResult = {
     optimized: null,
     rationale: null,
     reviewModel: "m",
-    rubricVersion: "content-review.v4",
+    rubricVersion: "content-review.v5",
     thresholdSnapshot: {},
     latencyMs: 12,
     degradedReason: null,

package/dist/index.d.ts

@@ -22,5 +22,29 @@ export declare function orchestrateVocAcquire(client: VocAcquireClient, params:
     exitCode: 0 | 1 | 2;
     output: Record<string, unknown>;
 }>;
+/**
+ * 业务失败(HTTP 200 + ok:false)的人类可读摘要：抓到了页面，但**不是正文**。
+ * 结论(outcome) + 阶段 + provider request id 全部打出来，别再让人对着通用错误猜。
+ */
+export declare function formatScrapeFailure(res: {
+    outcome: string;
+    finalUrl?: string;
+    targetStatusCode?: number | null;
+    failureStage?: string | null;
+    providerRequestId?: string | null;
+    requestId?: string | null;
+    attempts?: number;
+    retryAfterMs?: number | null;
+    classifierVersion?: string;
+    classificationEvidence?: Array<{
+        ruleId: string;
+        signal: string;
+    }>;
+}): string;
+/**
+ * 技术层失败(429/502/504 等)的摘要：把 provider typed code / 失败阶段 /
+ * request id 从 `details` 里挖出来，取代以前的「Scrape service returned HTTP 500」。
+ */
+export declare function formatScrapeError(e: unknown): string;
 export { program };
 //# sourceMappingURL=index.d.ts.map

package/dist/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":";AACA,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAuDpC,QAAA,MAAM,OAAO,SAAgB,CAAC;AA8+E9B,8DAA8D;AAC9D,wBAAgB,4BAA4B,CAC1C,IAAI,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAC5B,MAAM,GAAG,IAAI,CAOf;AAg+BD,MAAM,MAAM,gBAAgB,GAAG;IAC7B,QAAQ,CAAC,IAAI,EAAE,OAAO,GAAG,OAAO,CAAC,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC,CAAC;IAC1D,mBAAmB,CAAC,KAAK,EAAE,MAAM,GAAG,OAAO,CAAC;QAAE,MAAM,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;CACjE,CAAC;AAQF,wBAAsB,qBAAqB,CACzC,MAAM,EAAE,gBAAgB,EACxB,MAAM,EAAE;IACN,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IAC/B,YAAY,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IACtC,SAAS,EAAE,MAAM,CAAC;IAClB,SAAS,EAAE,MAAM,CAAC;IAClB,MAAM,EAAE,MAAM,CAAC;IACf,KAAK,EAAE,OAAO,CAAC;IACf,GAAG,CAAC,EAAE,MAAM,MAAM,CAAC;IACnB,KAAK,CAAC,EAAE,CAAC,EAAE,EAAE,MAAM,KAAK,OAAO,CAAC,IAAI,CAAC,CAAC;CACvC,GACA,OAAO,CAAC;IAAE,QAAQ,EAAE,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;IAAC,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAA;CAAE,CAAC,CAkJnE;AA47ED,OAAO,EAAE,OAAO,EAAE,CAAC"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":";AACA,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AA4DpC,QAAA,MAAM,OAAO,SAAgB,CAAC;AA2gF9B,8DAA8D;AAC9D,wBAAgB,4BAA4B,CAC1C,IAAI,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAC5B,MAAM,GAAG,IAAI,CAOf;AAg+BD,MAAM,MAAM,gBAAgB,GAAG;IAC7B,QAAQ,CAAC,IAAI,EAAE,OAAO,GAAG,OAAO,CAAC,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC,CAAC;IAC1D,mBAAmB,CAAC,KAAK,EAAE,MAAM,GAAG,OAAO,CAAC;QAAE,MAAM,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;CACjE,CAAC;AAQF,wBAAsB,qBAAqB,CACzC,MAAM,EAAE,gBAAgB,EACxB,MAAM,EAAE;IACN,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IAC/B,YAAY,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IACtC,SAAS,EAAE,MAAM,CAAC;IAClB,SAAS,EAAE,MAAM,CAAC;IAClB,MAAM,EAAE,MAAM,CAAC;IACf,KAAK,EAAE,OAAO,CAAC;IACf,GAAG,CAAC,EAAE,MAAM,MAAM,CAAC;IACnB,KAAK,CAAC,EAAE,CAAC,EAAE,EAAE,MAAM,KAAK,OAAO,CAAC,IAAI,CAAC,CAAC;CACvC,GACA,OAAO,CAAC;IAAE,QAAQ,EAAE,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;IAAC,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAA;CAAE,CAAC,CAkJnE;AAyqBD;;;GAGG;AACH,wBAAgB,mBAAmB,CAAC,GAAG,EAAE;IACvC,OAAO,EAAE,MAAM,CAAC;IAChB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,gBAAgB,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IACjC,YAAY,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IAC7B,iBAAiB,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IAClC,SAAS,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IAC1B,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,YAAY,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IAC7B,iBAAiB,CAAC,EAAE,MAAM,CAAC;IAC3B,sBAAsB,CAAC,EAAE,KAAK,CAAC;QAAE,MAAM,EAAE,MAAM,CAAC;QAAC,MAAM,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;CACpE,GAAG,MAAM,CAiBT;AAED;;;GAGG;AACH,wBAAgB,iBAAiB,CAAC,CAAC,EAAE,OAAO,GAAG,MAAM,CAyCpD;AAgyDD,OAAO,EAAE,OAAO,EAAE,CAAC"}

package/dist/index.js

@@ -8,6 +8,7 @@ import { registerSubredditStyleCommands } from "./register-subreddit-style.js";
 import { registerContentClassificationCommands } from "./register-content-classification.js";
 import { registerNotifyCommands } from "./register-notify.js";
 import { registerFeishuIdentityCommands } from "./register-feishu-identities.js";
+import { registerCalendarDirectiveCommands, calendarPatchNeedsGate, deliveryModeChangeRequiresApply, } from "./register-calendar-directives.js";
 import { assertApplyOrDryRun } from "./dry-run.js";
 import { withPermissionGate } from "./permission-runner.js";
 import { registerBatchAndExport } from "./register-batch.js";
@@ -284,6 +285,7 @@ dashboard
 });
 // --- calendar ---
 const calendar = program.command("calendar").description("发布日历条目");
+registerCalendarDirectiveCommands(calendar);
 calendar
     .command("list")
     .description("GET /calendar-entries")
@@ -311,6 +313,8 @@ calendar
     .option("-t, --team <teamId>", "Team UUID(缺省用当前 context)")
     .requiredOption("-e, --entry <entryId>", "Calendar entry UUID")
     .requiredOption("--body <json>", '例如 {"status":"succeeded","permalink":"https://www.reddit.com/..."}')
+    .option("--dry-run", "预览写入，不调用 API", false)
+    .option("--apply", "执行写入（仅 deliveryMode 变更必填）", false)
     .action(async (opts) => {
     let body;
     try {
@@ -320,7 +324,31 @@ calendar
         console.error("Invalid --body JSON");
         process.exit(1);
     }
-    const res = await requireClient().updateCalendarEntry(resolveTeamId(opts.team), opts.entry, body);
+    const teamId = resolveTeamId(opts.team);
+    // 两道门，判据不同，不能合成一条：
+    //
+    //   `--dry-run` **对任何 body 都必须生效** —— 全局 flag 契约是「预览写入，不调用
+    //   API」（docs/command-governance.md）。只在 deliveryMode 分支上认它，就会出现
+    //   「我明明加了 --dry-run，它却真的把状态改了」这种最不该有的意外写入。
+    //
+    //   `--apply` **只在 body 含 `deliveryMode` 时必填**。它是防重复发帖的唯一闸门
+    //   （方案 D1）：改它会让服务端在同一事务里撤销/创建内部 publish_post job 并
+    //   append 指令。`calendar set-delivery-mode` 已经带这道门，raw patch 必须同样
+    //   挡住，否则那道门只是装饰。其余字段（status/permalink/…）不加必填门，既有
+    //   自动化行为完全不变。
+    if (calendarPatchNeedsGate(body, opts) &&
+        !assertApplyOrDryRun(opts, {
+            command: deliveryModeChangeRequiresApply(body)
+                ? "calendar patch (deliveryMode)"
+                : "calendar patch",
+            danger: deliveryModeChangeRequiresApply(body)
+                ? "切换发布主体：转 external_agent 会撤销内部 publish_post job 并发 upsert 指令；转 hub_managed 会发 cancel 指令并建内部 job。改错 = 双发或谁都不发。"
+                : "更新日历条目",
+            summary: { teamId, calendarEntryId: opts.entry, body },
+        })) {
+        return;
+    }
+    const res = await requireClient().updateCalendarEntry(teamId, opts.entry, body);
     console.log(JSON.stringify(res, null, 2));
 });
 // --- reddit ---
@@ -3272,6 +3300,56 @@ intel
 const scrapeCmd = program
     .command("scrape")
     .description("通用网页抓取（Reddit .json 直连 / HTML 经 Firecrawl）");
+/**
+ * 业务失败(HTTP 200 + ok:false)的人类可读摘要：抓到了页面，但**不是正文**。
+ * 结论(outcome) + 阶段 + provider request id 全部打出来，别再让人对着通用错误猜。
+ */
+export function formatScrapeFailure(res) {
+    const evidence = (res.classificationEvidence ?? [])
+        .map((e) => `${e.ruleId}(${e.signal})`)
+        .join(", ");
+    const parts = [
+        `scrape 未取得正文：outcome=${res.outcome}`,
+        `targetStatus=${res.targetStatusCode ?? "null"}`,
+        `finalUrl=${res.finalUrl ?? "null"}`,
+        `stage=${res.failureStage ?? "null"}`,
+        `attempts=${res.attempts ?? 1}`,
+    ];
+    if (res.retryAfterMs != null)
+        parts.push(`retryAfterMs=${res.retryAfterMs}`);
+    parts.push(`requestId=${res.requestId ?? "null"}`);
+    parts.push(`providerRequestId=${res.providerRequestId ?? "null"}`);
+    if (res.classifierVersion)
+        parts.push(`classifier=${res.classifierVersion}`);
+    if (evidence)
+        parts.push(`evidence=[${evidence}]`);
+    return parts.join(" ");
+}
+/**
+ * 技术层失败(429/502/504 等)的摘要：把 provider typed code / 失败阶段 /
+ * request id 从 `details` 里挖出来，取代以前的「Scrape service returned HTTP 500」。
+ */
+export function formatScrapeError(e) {
+    const err = e;
+    const base = err?.message ?? String(e);
+    const d = err?.details;
+    if (!d?.outcome)
+        return base;
+    const p = d.diagnostics?.provider ?? {};
+    const x = d.diagnostics?.execution ?? {};
+    return [
+        base,
+        `code=${err.code ?? "null"}`,
+        `outcome=${d.outcome}`,
+        `stage=${x.failureStage ?? "null"}`,
+        `providerStatus=${p.providerStatusCode ?? "null"}`,
+        `providerErrorCode=${p.providerErrorCode ?? "null"}`,
+        `providerRequestId=${p.providerRequestId ?? "null"}`,
+        `providerScrapeId=${p.providerScrapeId ?? "null"}`,
+        `attempts=${x.attemptCount ?? 1}`,
+        ...(x.retryAfterMs != null ? [`retryAfterMs=${x.retryAfterMs}`] : []),
+    ].join(" ");
+}
 scrapeCmd
     .command("url <url>")
     .description("POST /v1/system/scrape —— 抓取任意网页(Reddit .json 或 HTML)")
@@ -3285,13 +3363,25 @@ scrapeCmd
             .map((s) => s.trim())
             .filter(Boolean)
         : undefined;
-    const res = await requireClient().scrapeUrl({
-        url,
-        formats,
-        mode: opts.redditJson ? "reddit-json" : undefined,
-        timeoutMs: opts.timeout ? Number(opts.timeout) : undefined,
-    });
-    console.log(JSON.stringify(res, null, 2));
+    try {
+        const res = await requireClient().scrapeUrl({
+            url,
+            formats,
+            mode: opts.redditJson ? "reddit-json" : undefined,
+            timeoutMs: opts.timeout ? Number(opts.timeout) : undefined,
+        });
+        console.log(JSON.stringify(res, null, 2));
+        if (!res.ok) {
+            // 业务失败(2xx 登录墙/验证码/空正文/目标站限流…):HTTP 是 200,但**不是正文**。
+            // 退出码必须非零,否则 shell/agent 会把「抓到登录页」当成抓取成功。
+            console.error(formatScrapeFailure(res));
+            process.exit(2);
+        }
+    }
+    catch (e) {
+        console.error(formatScrapeError(e));
+        process.exit(1);
+    }
 });
 scrapeCmd
     .command("status")