diff --git a/.github/workflows/quality.yml b/.github/workflows/quality.yml
index a8ba87f..185e516 100644
--- a/.github/workflows/quality.yml
+++ b/.github/workflows/quality.yml
@@ -23,5 +23,7 @@ jobs:
         run: npm run typecheck
       - name: Build
         run: npm run build
+      - name: Verify compiled adapters
+        run: npm run test:dist
       - name: Test
         run: npm test
diff --git a/README.en.md b/README.en.md
index 92da94c..f789ebe 100644
--- a/README.en.md
+++ b/README.en.md
@@ -17,7 +17,7 @@
 <p align="center">
   <a href="./LICENSE"><img src="https://img.shields.io/badge/License-AGPL--3.0-blue.svg?style=flat-square" alt="License: AGPL-3.0" /></a>
   <a href="https://www.npmjs.com/package/mancode"><img src="https://img.shields.io/npm/v/mancode?style=flat-square" alt="npm version" /></a>
-  <img src="https://img.shields.io/badge/status-V3%20beta%20v0.3.14-orange?style=flat-square" alt="Status: V3 beta v0.3.14" />
+  <img src="https://img.shields.io/badge/status-Continuity%20v0.3.16-2f855a?style=flat-square" alt="Status: mancode Continuity v0.3.16" />
   <img src="https://img.shields.io/badge/platforms-Claude%20Code%20%7C%20Cursor%20%7C%20Codex%20%7C%20Copilot%20%7C%20ZCode-5865F2?style=flat-square" alt="Platforms: Claude Code, Cursor, Codex in ChatGPT desktop and CLI, GitHub Copilot, ZCode" />
 </p>
 
@@ -34,6 +34,9 @@ different gears for different stakes: light solo mode for daily practice, `/man`
 for playoff-level engineering discipline, and coaching-staff subagents for
 research, planning, implementation, and review.
 
+**mancode Continuity** is the continuous-context runtime that carries tasks,
+decisions, and verification evidence safely into later conversations.
+
 [Installation](#installation) · [Usage](#usage)
 
 mancode ships with adapters for Claude Code, Cursor, Codex in the ChatGPT
@@ -102,7 +105,7 @@ explicitly resumes the existing TaskRef, preserving continuity without treating
 one window's temporary state as another window's identity.
 
 ```bash
-mancode status --json
+mancode status --brief --json
 mancode context session new --client claude-code
 mancode context resume <namespace:ULID> --session <id> --client claude-code
 mancode context show --purpose orient --session <id> --client claude-code
@@ -350,9 +353,8 @@ it should behave, and why previous decisions were made.
 
 ## Installation
 
-**Status**: V3 beta v0.3.14. Claude Code, Cursor, Codex in the ChatGPT desktop
-app and CLI, GitHub Copilot, and ZCode adapters are included. Stable release
-still requires the five-host real-session acceptance and the `context beta` B1 gate.
+**Status**: mancode Continuity v0.3.16. Claude Code, Cursor, Codex in the ChatGPT
+desktop app and CLI, GitHub Copilot, and ZCode adapters are included.
 
 Requires Node.js 20 or newer. macOS, Linux, Windows CMD, PowerShell, and Git Bash
 are supported. Git is optional: without it, initialization continues with solo
@@ -421,6 +423,7 @@ mancode init
 mancode init --legacy
 mancode status
 mancode status --json
+mancode status --brief --json
 mancode install <claude-code|cursor|codex|copilot|zcode>
 mancode list-platforms
 mancode team identity create --name "<name>"
@@ -449,7 +452,7 @@ mancode version
 Simplified output:
 
 ```text
-mancode v0.3.14
+mancode v0.3.16
 
 Project:     my-app
 Runtime:     ready
@@ -488,12 +491,15 @@ mancode init
 
 ### `mancode status`
 
-Shows activation, runtime binding, identity/session evidence, transport, and
-the physical readiness of each platform bootstrap and original mode entry.
+The default and full JSON views show activation, runtime binding,
+identity/session evidence, transport, and the physical readiness of each
+platform bootstrap and original mode entry. Coding agents should combine
+`--brief --json` to read only the compact Continuity runtime view.
 
 ```bash
 mancode status
 mancode status --json
+mancode status --brief --json
 ```
 
 ### `mancode workflow`
@@ -639,7 +645,7 @@ Ensure the `.cursor/rules/mancode-*.mdc` files exist. Rules with
 Mode-specific rules (manba, man, manteam, manps) trigger based on the
 description field — invoke them by asking for `/manba` or similar.
 
-### How to reinstall V3 adapters
+### How to reinstall Continuity adapters
 
 ```bash
 mancode uninstall claude-code --force
@@ -654,7 +660,7 @@ mancode install copilot
 mancode install zcode
 ```
 
-V3 authority is protected, so `mancode uninstall --all` does not delete workflow
+Continuity authority is protected, so `mancode uninstall --all` does not delete workflow
 authority. To inspect removable runtime records, run
 `mancode context compact --dry-run` first.
 
@@ -665,7 +671,7 @@ npm uninstall -g mancode
 ```
 
 Uninstalling each platform removes its mancode bootstrap while preserving
-user-authored rules, instructions, and V3 workflow data. The `--all` form is only
+user-authored rules, instructions, and Continuity workflow data. The `--all` form is only
 supported for projects explicitly initialized with `mancode init --legacy`.
 
 ## FAQ
diff --git a/README.md b/README.md
index a15a90e..f98f441 100644
--- a/README.md
+++ b/README.md
@@ -17,7 +17,7 @@
 <p align="center">
   <a href="./LICENSE"><img src="https://img.shields.io/badge/License-AGPL--3.0-blue.svg?style=flat-square" alt="许可证：AGPL-3.0" /></a>
   <a href="https://www.npmjs.com/package/mancode"><img src="https://img.shields.io/npm/v/mancode?style=flat-square" alt="npm 版本" /></a>
-  <img src="https://img.shields.io/badge/status-V3%20beta%20v0.3.14-orange?style=flat-square" alt="状态：V3 Beta v0.3.14" />
+  <img src="https://img.shields.io/badge/status-Continuity%20v0.3.16-2f855a?style=flat-square" alt="状态：mancode Continuity v0.3.16" />
   <img src="https://img.shields.io/badge/platforms-Claude%20Code%20%7C%20Cursor%20%7C%20Codex%20%7C%20Copilot%20%7C%20ZCode-5865F2?style=flat-square" alt="平台：Claude Code、Cursor、ChatGPT 桌面端 Codex、Codex CLI、GitHub Copilot、ZCode" />
 </p>
 
@@ -36,6 +36,8 @@
 日常任务用轻量 `solo`，关键任务用季后赛级别的 `/man`，复杂任务让教练组 subagents
 负责调研、计划、实现和审查。
 
+**mancode Continuity（连续上下文运行时）**负责把任务、决策和验证证据安全地带到后续对话。
+
 mancode 当前支持 Claude Code、Cursor、ChatGPT 桌面端中的 Codex、Codex CLI、
 GitHub Copilot 和 ZCode。每个平台继续使用原来的 `man*` 入口，并通过静态
 bootstrap 接入统一的 Context Pack 与 workflow authority。
@@ -92,7 +94,7 @@ mancode 把目标、需求、计划、检查结果和交接信息保存在稳定
 窗口的临时状态误当成另一个窗口的身份。
 
 ```bash
-mancode status --json
+mancode status --brief --json
 mancode context session new --client claude-code
 mancode context resume <namespace:ULID> --session <id> --client claude-code
 mancode context show --purpose orient --session <id> --client claude-code
@@ -305,9 +307,8 @@ src/components/
 
 ## 安装
 
-**状态**：V3 Beta v0.3.14。Claude Code、Cursor、ChatGPT 桌面端中的 Codex、
-Codex CLI 和 GitHub Copilot 均已接入；ZCode adapter 已接入。正式稳定发布仍需
-完成五平台真实宿主验收和 `context beta` B1 门禁。
+**状态**：mancode Continuity v0.3.16。Claude Code、Cursor、ChatGPT 桌面端中的
+Codex、Codex CLI、GitHub Copilot 和 ZCode adapter 均已接入。
 
 需要 Node.js 20 或更高版本。原生支持 macOS、Linux、Windows CMD、
 PowerShell 和 Git Bash。Git 是可选依赖：未安装时仍可初始化，只会把团队
@@ -374,6 +375,7 @@ mancode init
 mancode init --legacy
 mancode status
 mancode status --json
+mancode status --brief --json
 mancode install <claude-code|cursor|codex|copilot|zcode>
 mancode list-platforms
 mancode team identity create --name "<name>"
@@ -402,7 +404,7 @@ mancode version
 以下是简化输出示例：
 
 ```text
-mancode v0.3.14
+mancode v0.3.16
 
 Project:     my-app
 Runtime:     ready
@@ -440,12 +442,14 @@ mancode init
 
 ### `mancode status`
 
-显示 activation、runtime binding、identity/session evidence、transport 和各平台
-bootstrap/原 mode 入口的实际就绪状态。
+默认输出和完整 JSON 显示 activation、runtime binding、identity/session evidence、
+transport 和各平台 bootstrap/原 mode 入口的实际就绪状态。编码 Agent 应组合使用
+`--brief --json`，只读取精简的 Continuity 运行时视图。
 
 ```bash
 mancode status
 mancode status --json
+mancode status --brief --json
 ```
 
 ### `mancode workflow`
@@ -579,7 +583,7 @@ mancode/
 （context、practice、solo）在每次对话加载。模式规则（manba、man、manteam、
 manps）按 description 触发——输入 `/manba` 等关键词即可激活。
 
-### 如何重装 V3 适配器
+### 如何重装 Continuity 适配器
 
 ```bash
 mancode uninstall claude-code --force
@@ -594,7 +598,7 @@ mancode install copilot
 mancode install zcode
 ```
 
-V3 authority 受保护，`mancode uninstall --all` 不会删除工作流权威数据。需要
+Continuity authority 受保护，`mancode uninstall --all` 不会删除工作流权威数据。需要
 清理运行时保留记录时，先用 `mancode context compact --dry-run` 检查候选。
 
 ### 如何移除 CLI
@@ -604,7 +608,7 @@ npm uninstall -g mancode
 ```
 
 逐个平台卸载会移除对应的 mancode bootstrap，并保留用户自定义 rules、instructions
-和 V3 工作流数据。旧项目若明确使用 `mancode init --legacy`，才支持 legacy 的
+和 Continuity 工作流数据。旧项目若明确使用 `mancode init --legacy`，才支持 legacy 的
 `mancode uninstall --all --force`。
 
 ## 常见问题
diff --git a/package-lock.json b/package-lock.json
index af627fe..8ec0e32 100644
--- a/package-lock.json
+++ b/package-lock.json
@@ -1,12 +1,12 @@
 {
   "name": "mancode",
-  "version": "0.3.14",
+  "version": "0.3.16",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "mancode",
-      "version": "0.3.14",
+      "version": "0.3.16",
       "license": "AGPL-3.0-only",
       "dependencies": {
         "commander": "^12.1.0",
diff --git a/package.json b/package.json
index d5461a8..a5caea0 100644
--- a/package.json
+++ b/package.json
@@ -1,7 +1,7 @@
 {
   "name": "mancode",
-  "version": "0.3.14",
-  "description": "AI coding agent workflow harness. Five modes from practice to playoffs: stop over-engineering, reuse project context, and add multi-agent code review.",
+  "version": "0.3.16",
+  "description": "AI coding agent workflow harness with mancode Continuity for cross-conversation tasks, decisions, verification, and team coordination.",
   "type": "module",
   "license": "AGPL-3.0-only",
   "author": "whitelonng",
@@ -59,13 +59,14 @@
     "build": "tsup",
     "dev": "tsup --watch",
     "test": "vitest run",
+    "test:dist": "node scripts/verify-dist-adapters.mjs",
     "test:windows-smoke": "node scripts/windows-smoke.mjs",
     "test:watch": "vitest",
     "test:coverage": "vitest run --coverage",
     "lint": "biome check src tests",
     "typecheck": "tsc --noEmit",
     "format": "biome format --write src tests",
-    "prepublishOnly": "npm run lint && npm run typecheck && npm run build && npm test"
+    "prepublishOnly": "npm run lint && npm run typecheck && npm run build && npm run test:dist && npm test"
   },
   "dependencies": {
     "commander": "^12.1.0",
diff --git a/scripts/verify-dist-adapters.mjs b/scripts/verify-dist-adapters.mjs
new file mode 100644
index 0000000..7f2caac
--- /dev/null
+++ b/scripts/verify-dist-adapters.mjs
@@ -0,0 +1,92 @@
+import { spawnSync } from 'node:child_process';
+import { mkdir, mkdtemp, readFile, rm } from 'node:fs/promises';
+import { tmpdir } from 'node:os';
+import path from 'node:path';
+import process from 'node:process';
+import { fileURLToPath } from 'node:url';
+
+const repoRoot = path.resolve(
+  path.dirname(fileURLToPath(import.meta.url)),
+  '..',
+);
+const cliPath = path.join(repoRoot, 'dist', 'cli.js');
+const projectRoot = await mkdtemp(
+  path.join(tmpdir(), 'mancode-dist-adapters-'),
+);
+
+try {
+  await mkdir(projectRoot, { recursive: true });
+  runCli(['init', '--empty', '--platform', 'all', '--lang', 'en']);
+
+  const generatedFiles = [
+    '.claude/skills/mancode-v3/SKILL.md',
+    '.claude/skills/man/SKILL.md',
+    '.cursor/rules/mancode-v3.mdc',
+    '.cursor/commands/man.md',
+    'AGENTS.md',
+    '.agents/skills/man/SKILL.md',
+    '.github/copilot-instructions.md',
+    '.github/prompts/man.prompt.md',
+  ];
+
+  for (const relativePath of generatedFiles) {
+    const content = await readFile(path.join(projectRoot, relativePath), 'utf8');
+    assertGeneratedContract(relativePath, content);
+  }
+
+  const status = runCli(['status', '--brief', '--json']);
+  const parsed = JSON.parse(status.stdout);
+  assert(
+    parsed.runtime === 'mancode-continuity' && parsed.state === 'active',
+    `compiled status returned an unexpected runtime: ${status.stdout}`,
+  );
+  assert(
+    !status.stdout.includes('"v3"') && !status.stdout.includes('v3_active'),
+    `compiled brief status exposed internal version fields: ${status.stdout}`,
+  );
+
+  console.log(
+    `Verified ${generatedFiles.length} adapters generated by dist/cli.js.`,
+  );
+} finally {
+  await rm(projectRoot, { recursive: true, force: true });
+}
+
+function runCli(args) {
+  const result = spawnSync(process.execPath, [cliPath, ...args], {
+    cwd: projectRoot,
+    encoding: 'utf8',
+  });
+  assert(
+    result.status === 0,
+    `compiled CLI failed (${result.status}): ${result.stderr || result.stdout}`,
+  );
+  return result;
+}
+
+function assertGeneratedContract(relativePath, content) {
+  for (const required of [
+    'mancode status --brief --json',
+    'snapshot already obtained in this conversation',
+    'mancode Continuity',
+  ]) {
+    assert(
+      content.includes(required),
+      `${relativePath} is missing compiled guidance: ${required}`,
+    );
+  }
+  for (const forbidden of [
+    'mancode status --json',
+    'authority: "v3"',
+    'v3_active',
+  ]) {
+    assert(
+      !content.includes(forbidden),
+      `${relativePath} exposes stale guidance: ${forbidden}`,
+    );
+  }
+}
+
+function assert(condition, message) {
+  if (!condition) throw new Error(message);
+}
diff --git a/src/cli.ts b/src/cli.ts
index 41bff6b..2f1e0e2 100644
--- a/src/cli.ts
+++ b/src/cli.ts
@@ -112,6 +112,7 @@ program
   .command('status')
   .description('Show current mancode project status')
   .option('--json', 'Output as JSON (for scripts)')
+  .option('--brief', 'Output compact mancode Continuity runtime status')
   .action(async (options) => {
     const code = await status(process.cwd(), options);
     process.exitCode = code;
diff --git a/src/commands/context.ts b/src/commands/context.ts
index 6092124..9d64da9 100644
--- a/src/commands/context.ts
+++ b/src/commands/context.ts
@@ -55,6 +55,10 @@ import {
   resumeSession,
 } from '../runtime/session.js';
 import { readLocalActor } from '../team/actor.js';
+import {
+  capabilitiesFromGitRefCache,
+  readGitRefTeamCache,
+} from '../team/git-ref-cache.js';
 import {
   listGitRefWorkflowRepairs,
   recoverGitRefWorkflowRepair,
@@ -810,6 +814,16 @@ async function resolveContext(
     projectRoot: project.projectRoot,
     entityHomeStoreContext: project.runtime.entityHomeStoreContext,
   });
+  const capabilities =
+    project.project.config.transport.mode === 'git-ref'
+      ? capabilitiesFromGitRefCache(
+          project.project.config,
+          await readGitRefTeamCache(
+            project.projectRoot,
+            project.project.config,
+          ),
+        )
+      : undefined;
   return resolver.resolve({
     session,
     taskRef: request.taskRef,
@@ -823,6 +837,7 @@ async function resolveContext(
       adapterVersions: project.project.manifest.managedAdapters,
     },
     codeHead: await readCheckoutCodeHead(project.projectRoot),
+    ...(capabilities === undefined ? {} : { capabilities }),
   });
 }
 
diff --git a/src/commands/status.ts b/src/commands/status.ts
index a1cc529..499a0bb 100644
--- a/src/commands/status.ts
+++ b/src/commands/status.ts
@@ -81,6 +81,8 @@ export interface StatusState {
 export interface StatusOptions {
   /** --json: 输出 JSON（脚本用）*/
   json?: boolean;
+  /** --brief: 输出 Continuity 的紧凑运行时状态 */
+  brief?: boolean;
 }
 
 export interface StatusResult {
@@ -202,6 +204,19 @@ export interface V3StatusResult {
   legacyAuthorityPresent: boolean;
 }
 
+export interface ContinuityStatusResult {
+  schemaVersion: 1;
+  runtime: 'mancode-continuity';
+  state: 'active' | 'unavailable';
+  version: string;
+  project: string;
+  ready: boolean;
+  identity: V3StatusResult['localIdentity'];
+  session: V3StatusResult['currentSession'];
+  task: V3StatusResult['currentTask'];
+  blockers: string[];
+}
+
 interface DetectedTeamStatus {
   isTeam: boolean;
   contributors: number;
@@ -452,8 +467,11 @@ async function statusV3(
       adapters,
       legacyAuthorityPresent: legacy.authorityPresent,
     };
+    const output = options.brief ? toContinuityStatus(result) : result;
     if (options.json) {
-      console.log(JSON.stringify(result, null, 2));
+      console.log(JSON.stringify(output, null, 2));
+    } else if (options.brief) {
+      printContinuityText(output as ContinuityStatusResult);
     } else {
       printV3Text(result);
     }
@@ -466,6 +484,30 @@ async function statusV3(
   }
 }
 
+function toContinuityStatus(result: V3StatusResult): ContinuityStatusResult {
+  const blockers = [
+    ...result.compatibility.failures,
+    ...(result.runtime.error === null ? [] : [result.runtime.error]),
+  ];
+  const ready =
+    result.activation.state === 'v3_active' &&
+    result.compatibility.readAllowed &&
+    result.compatibility.writeAllowed &&
+    result.runtime.binding === 'ready';
+  return {
+    schemaVersion: 1,
+    runtime: 'mancode-continuity',
+    state: ready ? 'active' : 'unavailable',
+    version: result.version,
+    project: result.project,
+    ready,
+    identity: result.localIdentity,
+    session: result.currentSession,
+    task: result.currentTask,
+    blockers,
+  };
+}
+
 /** Status is read-only: an invalid ambient session is simply not current. */
 async function readV3StatusSession(rootDir: string) {
   const sessionId = process.env.MANCODE_SESSION_ID;
@@ -846,7 +888,7 @@ function printText(r: StatusResult): void {
 
 function printV3Text(result: V3StatusResult): void {
   console.log('');
-  console.log(`mancode v${result.version}`);
+  console.log(`mancode Continuity v${result.version}`);
   console.log('');
   console.log(`Project:     ${result.project}`);
   console.log(
@@ -913,6 +955,16 @@ function printV3Text(result: V3StatusResult): void {
   console.log('');
 }
 
+function printContinuityText(result: ContinuityStatusResult): void {
+  console.log(`mancode Continuity v${result.version}`);
+  console.log(`Project:  ${result.project}`);
+  console.log(`State:    ${result.state}`);
+  console.log(`Ready:    ${result.ready ? 'yes' : 'no'}`);
+  if (result.blockers.length > 0) {
+    console.log(`Blockers: ${result.blockers.join('; ')}`);
+  }
+}
+
 function formatActivationState(state: string): string {
   return state.replace(/^v3_/, '').replaceAll('_', ' ');
 }
diff --git a/src/commands/uninstall.ts b/src/commands/uninstall.ts
index 4fa4a97..34e1cb4 100644
--- a/src/commands/uninstall.ts
+++ b/src/commands/uninstall.ts
@@ -104,7 +104,7 @@ async function uninstallV3(
   if (!platform || options.all) {
     console.error('✗  mancode authority is protected from bulk uninstall.');
     console.error(
-      '   Remove a single bootstrap with `mancode uninstall <platform>`. Inspect runtime retention with `mancode context compact --dry-run`; V3 workflow authority is not bulk-deleted.',
+      '   Remove a single bootstrap with `mancode uninstall <platform>`. Inspect runtime retention with `mancode context compact --dry-run`; Continuity workflow authority is not bulk-deleted.',
     );
     return EXIT_V3_AUTHORITY_PROTECTED;
   }
diff --git a/src/commands/workflow.ts b/src/commands/workflow.ts
index e98762b..cc8c133 100644
--- a/src/commands/workflow.ts
+++ b/src/commands/workflow.ts
@@ -384,7 +384,7 @@ function workflowCleanV3(options: WorkflowOptions): number {
   return printV3Error(
     options.json,
     'MANCODE_V3_WORKFLOW_CLEAN_UNSUPPORTED',
-    'V3 workflow authority is durable and is not deleted by workflow clean. Use `mancode context compact --dry-run` to inspect eligible runtime retention records.',
+    'Continuity workflow authority is durable and is not deleted by workflow clean. Use `mancode context compact --dry-run` to inspect eligible runtime retention records.',
     EXIT_INVALID_ARG,
   );
 }
diff --git a/src/context/task-head-reconcile.ts b/src/context/task-head-reconcile.ts
index 78d5201..28b1a92 100644
--- a/src/context/task-head-reconcile.ts
+++ b/src/context/task-head-reconcile.ts
@@ -31,15 +31,18 @@ import {
 } from '../runtime/task-operation.js';
 import type { ClaimV1 } from '../team/claims.js';
 import { deriveClaimValidity } from '../team/conflicts.js';
+import { createGitRefTeamManifestStore } from '../team/git-ref-client.js';
 import type { HandoffV1 } from '../team/handoff.js';
+import type { ProjectConfigV1 } from '../team/policy.js';
 import {
   type TaskAggregateManifestV1,
   taskAggregateDigest,
 } from './aggregate.js';
 import { type Ulid, assertUlid, createUlid } from './ids.js';
+import type { SchemaManifestV1 } from './manifest.js';
 import { type StoredTaskSnapshot, V3ContextStore } from './store.js';
 import { assertTaskCodeHeadUnchanged } from './task-mutation.js';
-import { type TaskRef, parseTaskRefValue } from './task-ref.js';
+import { type TaskRef, parseTaskRefValue, sameTaskRef } from './task-ref.js';
 
 const execFile = promisify(execFileCallback);
 
@@ -119,13 +122,19 @@ export async function previewV3TaskHeadReconcile(
     input.projectRoot,
     task.location.taskRoot,
   );
+  const project = await store.readProjectSnapshot();
+  await assertRemoteTaskUnpublished({
+    projectRoot: input.projectRoot,
+    taskRef,
+    config: project.config,
+    manifest: project.manifest,
+  });
   assertReconcileEligible({
     task,
     aggregate,
     previousFence,
     codeHead,
     actorId: input.sessionActorId,
-    transportMode: (await store.readProjectSnapshot()).config.transport.mode,
     claims: coordination.claims,
     handoffs: coordination.handoffs,
     now,
@@ -194,13 +203,18 @@ export async function reconcileV3TaskHead(
     if (reread.fingerprint !== context.task.fingerprint) {
       throw new Error('MANCODE_CONTEXT_STALE');
     }
+    await assertRemoteTaskUnpublished({
+      projectRoot: context.projectRoot,
+      taskRef,
+      config: context.project.config,
+      manifest: context.project.manifest,
+    });
     assertReconcileEligible({
       task: context.task,
       aggregate,
       previousFence,
       codeHead,
       actorId: context.session.actorId,
-      transportMode: context.project.config.transport.mode,
       claims: context.coordination.claims,
       handoffs: context.coordination.handoffs,
       now: context.now,
@@ -396,14 +410,10 @@ function assertReconcileEligible(input: {
   previousFence: TaskHeadFenceV1;
   codeHead: string;
   actorId: Ulid;
-  transportMode: 'local' | 'git-ref';
   claims: ClaimV1[];
   handoffs: HandoffV1[];
   now: Date;
 }): void {
-  if (input.transportMode !== 'local') {
-    throw new Error('MANCODE_GIT_REF_TRANSPORT_NOT_IMPLEMENTED');
-  }
   if (input.task.metadata.ownerActorId !== input.actorId) {
     throw new Error('MANCODE_TASK_OWNER_REQUIRED');
   }
@@ -420,6 +430,31 @@ function assertReconcileEligible(input: {
   assertHandoffsConsistent(input.handoffs);
 }
 
+async function assertRemoteTaskUnpublished(input: {
+  projectRoot: string;
+  taskRef: TaskRef;
+  config: ProjectConfigV1;
+  manifest: SchemaManifestV1;
+}): Promise<void> {
+  if (input.config.transport.mode !== 'git-ref') return;
+  const snapshot = await createGitRefTeamManifestStore(
+    input.projectRoot,
+    input.config,
+    input.manifest,
+  ).pull();
+  const remote = snapshot.manifest;
+  if (remote === null) return;
+  const hasTaskAuthority = [
+    ...remote.ownershipFences,
+    ...remote.taskBundles,
+    ...remote.claims,
+    ...remote.handoffs,
+  ].some((entity) => sameTaskRef(entity.taskRef, input.taskRef));
+  if (hasTaskAuthority) {
+    throw new Error('MANCODE_GIT_REF_TASK_ALREADY_PUBLISHED');
+  }
+}
+
 function assertClaimsConsistent(
   claims: ClaimV1[],
   task: StoredTaskSnapshot,
diff --git a/src/context/workflow-create.ts b/src/context/workflow-create.ts
index 76be0b7..558d440 100644
--- a/src/context/workflow-create.ts
+++ b/src/context/workflow-create.ts
@@ -51,6 +51,11 @@ import { createTaskHeadFence } from '../runtime/task-head-store.js';
 import { readSharedActorProfile } from '../team/actor.js';
 import type { TeamAssessment } from '../team/assessment.js';
 import { createAuthorizationBasis } from '../team/authorization.js';
+import {
+  capabilitiesFromGitRefCache,
+  readGitRefTeamCache,
+} from '../team/git-ref-cache.js';
+import type { ProjectConfigV1 } from '../team/policy.js';
 import { assertTransportCoordinationWriteAllowed } from '../team/transport-migration-freeze.js';
 import { capabilitiesFromProjectConfig } from '../team/transport.js';
 import { VERSION } from '../version.js';
@@ -225,6 +230,7 @@ export async function createV3Workflow(
   if (taskRef.namespace === 'shared') {
     await assertSharedCreationPrerequisites(
       projectRoot,
+      project.config,
       session.actorId,
       input.sharedPrivacyConfirmed === true,
       participants,
@@ -247,6 +253,11 @@ export async function createV3Workflow(
     parent,
     participants,
     explicitScope: input.implementationScope,
+    ownershipEpoch:
+      taskRef.namespace === 'shared' &&
+      project.config.transport.mode === 'git-ref'
+        ? 0
+        : 1,
     timestamp,
   });
   const homeStore = resolveTaskEntityHomeStore(
@@ -527,6 +538,7 @@ async function resolveParentCreationContext(
 
 async function assertSharedCreationPrerequisites(
   projectRoot: string,
+  config: ProjectConfigV1,
   actorId: Ulid,
   privacyConfirmed: boolean,
   participants: readonly Ulid[],
@@ -534,19 +546,40 @@ async function assertSharedCreationPrerequisites(
   if (!privacyConfirmed) {
     throw new Error('MANCODE_PRIVACY_CONFIRMATION_REQUIRED');
   }
-  if ((await readSharedActorProfile(projectRoot, actorId)) === null) {
+  if (!(await actorHasSharedProfile(projectRoot, config, actorId))) {
     throw new Error('MANCODE_JOIN_REQUIRED');
   }
   for (const participantActorId of participants) {
     if (participantActorId === actorId) continue;
     if (
-      (await readSharedActorProfile(projectRoot, participantActorId)) === null
+      !(await actorHasSharedProfile(projectRoot, config, participantActorId))
     ) {
       throw new Error('MANCODE_PARTICIPANT_JOIN_REQUIRED');
     }
   }
 }
 
+async function actorHasSharedProfile(
+  projectRoot: string,
+  config: ProjectConfigV1,
+  actorId: Ulid,
+): Promise<boolean> {
+  if ((await readSharedActorProfile(projectRoot, actorId)) !== null)
+    return true;
+  if (config.transport.mode !== 'git-ref') return false;
+  const cache = await readGitRefTeamCache(projectRoot, config);
+  if (
+    capabilitiesFromGitRefCache(config, cache).transportFreshness !== 'fresh'
+  ) {
+    throw new Error('MANCODE_TRANSPORT_NOT_FRESH');
+  }
+  return (
+    cache?.manifest?.actorProfiles.some(
+      (profile) => profile.actorId === actorId,
+    ) === true
+  );
+}
+
 async function requireSharedCodeHead(projectRoot: string): Promise<string> {
   const head = await readCheckoutCodeHead(projectRoot);
   if (head === null) throw new Error('MANCODE_TASK_HEAD_CODE_REF_UNAVAILABLE');
@@ -595,6 +628,7 @@ function buildInitialEntities(input: {
   parent: ParentCreationContext | null;
   participants: Ulid[];
   explicitScope: WorkflowCreateScope | undefined;
+  ownershipEpoch: number;
   timestamp: string;
 }): InitialEntities {
   const scope = initialScope(input.parent, input.explicitScope);
@@ -691,7 +725,7 @@ function buildInitialEntities(input: {
     transitionState: 'stable',
     lastOperationId: input.operationId,
     ownerActorId: input.actorId,
-    ownershipEpoch: 1,
+    ownershipEpoch: input.ownershipEpoch,
     participants: input.participants,
     createdBy: {
       actorId: input.actorId,
diff --git a/src/installers/v3-adapter.ts b/src/installers/v3-adapter.ts
index 3de59c4..c9b2cce 100644
--- a/src/installers/v3-adapter.ts
+++ b/src/installers/v3-adapter.ts
@@ -581,8 +581,8 @@ export function renderV3Bootstrap(platform: PlatformName): string {
   const sessionArguments = sessionArgumentsFor(platform);
   const sessionCreationGuidance =
     platform === 'codex' || platform === 'zcode'
-      ? 'When status has no `currentSession`, first reuse any explicit session ID already returned in this conversation. Only when neither exists, create one once with `mancode context session new --client codex` in Codex or `mancode context session new --client zcode` in ZCode.'
-      : `When status has no \`currentSession\`, first reuse any explicit session ID already returned in this conversation. Only when neither exists, create one once with \`mancode context session new --client ${platform}\`.`;
+      ? 'When status has no `session`, first reuse any explicit session ID already returned in this conversation. Only when neither exists, create one once with `mancode context session new --client codex` in Codex or `mancode context session new --client zcode` in ZCode.'
+      : `When status has no \`session\`, first reuse any explicit session ID already returned in this conversation. Only when neither exists, create one once with \`mancode context session new --client ${platform}\`.`;
   const spikePlatform =
     platform === 'codex' || platform === 'zcode'
       ? 'the active Codex or ZCode host'
@@ -597,12 +597,12 @@ export function renderV3Bootstrap(platform: PlatformName): string {
     '',
     `- Platform: ${platformLabel}. This file is a non-authoritative bootstrap.`,
     '- Locate the project root before running mancode commands.',
-    '- First run `mancode status --json` from the project root.',
-    '- Treat status values such as `authority: "v3"` and `v3_active` as internal machine fields. In operator-facing narration, say `mancode`; never prefix a mode or action with a version label.',
+    '- Reuse a `mancode status --brief --json` snapshot already obtained in this conversation. Only when no such snapshot exists, run it once from the project root.',
+    '- The compact status is the public mancode Continuity runtime view. In operator-facing narration, say `mancode` or `mancode Continuity`; never prefix a mode or action with a version label.',
     '- An explicitly invoked original `man`, `manba`, `manteam`, `manps`, or `mansolo` entry supplies its authorized action. Its mode-specific steps override conflicting generic no-task or mutation guidance below.',
     '- In particular, `manps` may run local health scans without an actor, session, or TaskRef. `mansolo` needs them only for an explicit governed handoff.',
-    '- If status has no `localIdentity.actorId`, ask for a display name and run `mancode team identity create --name "<display name>"` before creating a session.',
-    '- If status reports `currentSession`, reuse it. `currentTask: null` and `MANCODE_TASK_REQUIRED` do not make a session stale.',
+    '- If status has no `identity.actorId`, ask for a display name and run `mancode team identity create --name "<display name>"` before creating a session.',
+    '- If status reports `session`, reuse it. `task: null` and `MANCODE_TASK_REQUIRED` do not make a session stale.',
     `- ${sessionCreationGuidance} Pass its returned \`sessionId\` and matching client as \`${sessionArguments}\` to every later session command; an \`export\` inside one command tool does not persist to later command tools.`,
     '- Outside an invoked original mode entry, if no current task and no task is explicitly supplied, report "no task bound" and stop. Do not probe workflow subcommands to work around `MANCODE_TASK_REQUIRED`.',
     '- Bootstrap discovery is read-only: before the operator explicitly requests task work, do not run `mancode init`, `mancode migrate`, `mancode workflow`, or inspect mancode installed package/source.',
@@ -627,6 +627,8 @@ export function renderV3ModeEntry(
   const definition = V3_MODE_DEFINITIONS[mode];
   const sessionArguments = sessionArgumentsFor(platform);
   const sessionClientGuidance = sessionClientGuidanceFor(platform);
+  const statusGuidance =
+    'Reuse a `mancode status --brief --json` snapshot already obtained in this conversation. Only when none exists, run it once from the project root.';
   const sessionCreationGuidance =
     platform === 'codex' || platform === 'zcode'
       ? 'If status has no current session, reuse an explicit session ID already retained in this conversation. Only if neither exists, run `mancode context session new --client codex` in Codex or `mancode context session new --client zcode` in ZCode exactly once, then retain the returned session ID.'
@@ -634,31 +636,31 @@ export function renderV3ModeEntry(
   let authoritySteps: string[];
   if (mode === 'manps') {
     authoritySteps = [
-      '1. Run `mancode status --json` from the project root and require active mancode authority.',
+      `1. ${statusGuidance} Require an active, ready mancode Continuity runtime.`,
       '2. Run the health action below directly. A local scan needs no TaskRef, actor identity, or explicit session.',
       '3. Never read or write legacy mode authority before or after the scan.',
     ];
   } else if (mode === 'mansolo') {
     authoritySteps = [
-      '1. Run `mancode status --json` from the project root. Never read or write legacy mode authority.',
+      `1. ${statusGuidance} Never read or write legacy mode authority.`,
       '2. If no governed task is being handed off, continue with focused solo work without creating a persistent mode, actor, session, or TaskRef.',
-      `3. For an explicit governed handoff, ensure \`localIdentity.actorId\`, reuse or create the current session, and bind the existing TaskRef. ${sessionCreationGuidance} ${sessionClientGuidance}`,
+      `3. For an explicit governed handoff, ensure \`identity.actorId\`, reuse or create the current session, and bind the existing TaskRef. ${sessionCreationGuidance} ${sessionClientGuidance}`,
       `4. For that governed task only, read \`mancode context show --purpose implement ${sessionArguments}\` using the bound or explicit TaskRef.`,
     ];
   } else if (mode === 'man') {
     authoritySteps = [
-      '1. Run `mancode status --json` from the project root. Never read or write the legacy authority file.',
+      `1. ${statusGuidance} Never read or write the legacy authority file.`,
       '2. If the request is only a read-only project orientation, introduction, explanation, or summary, inspect the repository and answer directly without creating an actor, session, TaskRef, or workflow. Do not turn an orientation request into a `plan_only` workflow.',
-      '3. Only for requested governed task work, if `localIdentity.actorId` is absent, ask for a display name and run `mancode team identity create --name "<display name>"`.',
-      `4. Reuse \`currentSession.sessionId\` when present. ${sessionCreationGuidance} ${sessionClientGuidance}`,
+      '3. Only for requested governed task work, if `identity.actorId` is absent, ask for a display name and run `mancode team identity create --name "<display name>"`.',
+      `4. Reuse \`session.sessionId\` when present. ${sessionCreationGuidance} ${sessionClientGuidance}`,
       `5. Reuse the current TaskRef. To bind a supplied existing task, run \`mancode context resume <namespace:ULID> ${sessionArguments}\`.`,
       `6. For an existing task, read only the needed Context Pack with \`mancode context show --purpose ${definition.contextPurpose} ${sessionArguments}\`; include \`--task <namespace:ULID>\` when it is not yet bound. For a new task, create it through the mode action first, then read the returned TaskRef's Context Pack.`,
     ];
   } else {
     authoritySteps = [
-      '1. Run `mancode status --json` from the project root. Never read or write the legacy authority file.',
-      '2. If `localIdentity.actorId` is absent, ask for a display name and run `mancode team identity create --name "<display name>"`.',
-      `3. Reuse \`currentSession.sessionId\` when present. ${sessionCreationGuidance} ${sessionClientGuidance}`,
+      `1. ${statusGuidance} Never read or write the legacy authority file.`,
+      '2. If `identity.actorId` is absent, ask for a display name and run `mancode team identity create --name "<display name>"`.',
+      `3. Reuse \`session.sessionId\` when present. ${sessionCreationGuidance} ${sessionClientGuidance}`,
       `4. Reuse the current TaskRef. To bind a supplied existing task, run \`mancode context resume <namespace:ULID> ${sessionArguments}\`.`,
       `5. For an existing task, read only the needed Context Pack with \`mancode context show --purpose ${definition.contextPurpose} ${sessionArguments}\`; include \`--task <namespace:ULID>\` when it is not yet bound. For a new task, create it through the mode action first, then read the returned TaskRef's Context Pack.`,
     ];
@@ -700,7 +702,7 @@ export function renderV3ModeEntry(
     '',
     ...actions,
     '',
-    'Treat status values such as `authority: "v3"` and `v3_active` as internal machine fields. In operator-facing narration, say `mancode`; never prefix this mode or its actions with a version label.',
+    'Treat the compact status as the public mancode Continuity runtime view. In operator-facing narration, say `mancode` or `mancode Continuity`; never prefix this mode or its actions with a version label.',
     '',
     mutationGuidance,
     '',
@@ -1313,7 +1315,7 @@ function renderRetiredCursorRule(target: V3LegacyAdapterFileTarget): string {
           : legacyName;
   const guidance =
     mode === 'context' || mode === 'practice'
-      ? 'Run `mancode status --json`, then use the matching mancode mode command.'
+      ? 'Run `mancode status --brief --json`, then use the matching mancode mode command.'
       : `Use the \`/${mode}\` mancode mode command.`;
   return [
     '---',
diff --git a/src/team/git-ref-operation.ts b/src/team/git-ref-operation.ts
index a2dcb04..0739f2c 100644
--- a/src/team/git-ref-operation.ts
+++ b/src/team/git-ref-operation.ts
@@ -13,6 +13,7 @@ import {
   sameTaskRef,
 } from '../context/task-ref.js';
 import { parseWorkflowMetadata } from '../context/workflow-metadata.js';
+import { readOperationJournal } from '../runtime/operation-store.js';
 import { readCheckoutBranch } from '../runtime/project-runtime.js';
 import {
   type OpenedV3TaskOperation,
@@ -203,7 +204,6 @@ export async function syncGitRefTask(
     now,
   });
   try {
-    await assertCleanGitWorktree(context.projectRoot);
     const transport = createGitRefTeamManifestStore(
       context.projectRoot,
       context.project.config,
@@ -211,6 +211,21 @@ export async function syncGitRefTask(
     );
     const snapshot = await transport.pull();
     const bundle = await bundleFromContext(context, now);
+    const remoteTaskPublished =
+      snapshot.manifest?.ownershipFences.some((fence) =>
+        sameTaskRef(fence.taskRef, taskRef),
+      ) === true &&
+      snapshot.manifest.taskBundles.some((remoteBundle) =>
+        sameTaskRef(remoteBundle.taskRef, taskRef),
+      );
+    const taskAuthorityChangesAllowed =
+      remoteTaskPublished &&
+      (remoteBundleMatches(snapshot, bundle) ||
+        (await hasCommittedHandoffCheckpoint(context)));
+    await assertCleanGitWorktree(
+      context.projectRoot,
+      taskAuthorityChangesAllowed ? taskRef : undefined,
+    );
     const result = await synchronizeBundle(
       context,
       transport,
@@ -471,7 +486,7 @@ export async function createGitRefHandoffDraft(
     assertContextMatchesRemoteBundle(context, bundle);
     await Promise.all([
       assertGitRefBundleCodeReachable(context.projectRoot, bundle),
-      assertCleanGitWorktree(context.projectRoot),
+      assertCleanGitWorktree(context.projectRoot, taskRef),
     ]);
     const checkpoint = context.task.latestCheckpoint;
     if (
@@ -571,7 +586,10 @@ export async function mutateGitRefHandoff(
   });
   try {
     if (input.mutation.kind === 'offer') {
-      await assertCleanGitWorktree(opened.context.projectRoot);
+      await assertCleanGitWorktree(
+        opened.context.projectRoot,
+        opened.handoff.taskRef,
+      );
     }
     const kind =
       input.mutation.kind === 'offer'
@@ -848,15 +866,49 @@ function defaultRemoteHandoffSummary(checkpoint: CheckpointV1): HandoffSummary {
   };
 }
 
-async function assertCleanGitWorktree(projectRoot: string): Promise<void> {
-  const { stdout } = await execFile(
-    'git',
-    ['status', '--porcelain=v1', '--untracked-files=all'],
-    { cwd: projectRoot, windowsHide: true },
-  );
+async function assertCleanGitWorktree(
+  projectRoot: string,
+  materializedTaskRef?: TaskRef,
+): Promise<void> {
+  const args = ['status', '--porcelain=v1', '--untracked-files=all'];
+  if (materializedTaskRef !== undefined) {
+    args.push(
+      '--',
+      '.',
+      `:(exclude,top).mancode/shared/workflows/${materializedTaskRef.taskId}/**`,
+    );
+  }
+  const { stdout } = await execFile('git', args, {
+    cwd: projectRoot,
+    windowsHide: true,
+  });
   if (stdout.trim()) throw new Error('MANCODE_HANDOFF_DIRTY_WORKTREE');
 }
 
+async function hasCommittedHandoffCheckpoint(
+  context: OpenedV3TaskOperation,
+): Promise<boolean> {
+  const checkpoint = context.task.latestCheckpoint;
+  if (
+    checkpoint === null ||
+    checkpoint.kind !== 'handoff_offered' ||
+    context.task.metadata.lastOperationId !== checkpoint.operationId ||
+    context.task.aggregate?.latestCheckpointId !== checkpoint.checkpointId
+  ) {
+    return false;
+  }
+  const journal = await readOperationJournal(
+    context.homeStore,
+    checkpoint.operationId,
+  );
+  return (
+    journal?.type === 'checkpoint_create' &&
+    journal.state === 'committed' &&
+    journal.actorId === context.session.actorId &&
+    journal.sessionId === context.session.sessionId
+  );
+}
+
 function assertPositiveRevision(value: number, label: string): void {
   if (!Number.isSafeInteger(value) || value < 1) {
     throw new Error(`${label} must be a positive integer`);
@@ -883,10 +935,7 @@ async function synchronizeBundle(
   const fence = manifest.ownershipFences.find((candidate) =>
     sameTaskRef(candidate.taskRef, context.taskRef),
   );
-  if (fence === undefined) {
-    throw new Error('MANCODE_TRANSPORT_MIGRATION_REQUIRED');
-  }
-  if (remoteBundleMatches(snapshot, bundle)) {
+  if (fence !== undefined && remoteBundleMatches(snapshot, bundle)) {
     return {
       remoteRevision: manifest.revision,
       ownershipEpoch: fence.ownershipEpoch,
@@ -900,7 +949,7 @@ async function synchronizeBundle(
     actorId: context.session.actorId,
     taskRef: context.taskRef,
     expectedRemoteRevision: manifest.revision,
-    expectedOwnershipEpoch: fence.ownershipEpoch,
+    expectedOwnershipEpoch: fence?.ownershipEpoch ?? 0,
     taskBundle: bundle,
     now,
   });
diff --git a/tests/git-ref-cross-clone-e2e.test.ts b/tests/git-ref-cross-clone-e2e.test.ts
index 7816274..73e4902 100644
--- a/tests/git-ref-cross-clone-e2e.test.ts
+++ b/tests/git-ref-cross-clone-e2e.test.ts
@@ -4,7 +4,11 @@ import os from 'node:os';
 import path from 'node:path';
 import { promisify } from 'node:util';
 import { afterEach, describe, expect, it } from 'vitest';
-import { teamSyncPull } from '../src/commands/team.js';
+import {
+  contextReconcileTaskHead,
+  contextResume,
+} from '../src/commands/context.js';
+import { teamClaim, teamSyncPull, teamSyncPush } from '../src/commands/team.js';
 import { initializeV3Project } from '../src/commands/v3-init.js';
 import { workflow } from '../src/commands/workflow.js';
 import { createV3Checkpoint } from '../src/context/checkpoint-create.js';
@@ -78,6 +82,238 @@ afterEach(async () => {
 });
 
 describe('git-ref coordination across independent clones', () => {
+  it('publishes a newly created git-ref task after its Git commit', async () => {
+    const fixture = await createFixture();
+    const storeA = await strictStore(fixture.cloneA);
+    const [profileA, profileB] = fixture.profiles;
+    if (profileA === undefined || profileB === undefined) {
+      throw new Error('missing fixture actor profiles');
+    }
+    await storeA.publishActorProfile({
+      operationId: id(15),
+      expectedRemoteRevision: 0,
+      profile: profileA,
+    });
+    await storeA.publishActorProfile({
+      operationId: id(16),
+      expectedRemoteRevision: 1,
+      profile: profileB,
+    });
+    await expect(
+      captureJson(() => teamSyncPull(fixture.cloneA, { json: true })),
+    ).resolves.toMatchObject({
+      exitCode: 0,
+      value: { remoteRevision: 2 },
+    });
+    await rm(
+      path.join(
+        fixture.cloneA,
+        '.mancode',
+        'shared',
+        'team',
+        'actors',
+        `${ACTOR_B}.json`,
+      ),
+    );
+
+    const created = await captureJson<WorkflowCreateJson>(() =>
+      workflow(
+        fixture.cloneA,
+        'create',
+        [
+          'manteam',
+          'Publish a new shared task through the public sync sequence.',
+        ],
+        {
+          session: SESSION_A,
+          client: 'vitest-a',
+          participants: [ACTOR_B],
+          confirmShared: true,
+          json: true,
+        },
+      ),
+    );
+    expect(created).toMatchObject({
+      exitCode: 0,
+      value: { metadata: { revision: 1, ownershipEpoch: 0 } },
+    });
+    await publishSharedActorProfile(fixture.cloneA, profileB);
+    const taskRef = created.value.taskRef;
+    const formattedTaskRef = formatTaskRef(taskRef);
+
+    await writeFile(
+      path.join(fixture.cloneA, 'README.md'),
+      '# git-ref cross-clone fixture\n\ncode head\n\nteam task\n',
+    );
+    const dirtyPush = await captureJson<CommandJson>(() =>
+      teamSyncPush(fixture.cloneA, {
+        task: formattedTaskRef,
+        expectedTaskRevision: '1',
+        session: SESSION_A,
+        client: 'vitest-a',
+        json: true,
+      }),
+    );
+    expect(dirtyPush).toMatchObject({
+      exitCode: 3,
+      value: { error: { code: 'MANCODE_HANDOFF_DIRTY_WORKTREE' } },
+    });
+
+    const taskRoot = path.join(
+      '.mancode',
+      'shared',
+      'workflows',
+      taskRef.taskId,
+    );
+    await git(fixture.cloneA, ['add', '--force', 'README.md', taskRoot]);
+    await git(fixture.cloneA, ['commit', '-m', 'publish shared task']);
+    const reconciled = await captureJson<CommandJson>(() =>
+      contextReconcileTaskHead(fixture.cloneA, formattedTaskRef, {
+        expectedFenceRevision: '1',
+        fromGit: true,
+        session: SESSION_A,
+        client: 'vitest-a',
+        json: true,
+      }),
+    );
+    expect(reconciled.exitCode).toBe(0);
+
+    const synced = await captureJson<CommandJson>(() =>
+      teamSyncPush(fixture.cloneA, {
+        task: formattedTaskRef,
+        expectedTaskRevision: '1',
+        session: SESSION_A,
+        client: 'vitest-a',
+        json: true,
+      }),
+    );
+    expect(synced).toMatchObject({
+      exitCode: 0,
+      value: {
+        changed: true,
+        remoteRevision: 3,
+        ownershipEpoch: 0,
+        taskRevision: 1,
+      },
+    });
+    const publishedReconcile = await captureJson<CommandJson>(() =>
+      contextReconcileTaskHead(fixture.cloneA, formattedTaskRef, {
+        expectedFenceRevision: '2',
+        fromGit: true,
+        session: SESSION_A,
+        client: 'vitest-a',
+        json: true,
+      }),
+    );
+    expect(publishedReconcile).toMatchObject({
+      exitCode: 3,
+      value: { error: { code: 'MANCODE_GIT_REF_TASK_ALREADY_PUBLISHED' } },
+    });
+
+    await git(fixture.cloneA, ['push', 'origin', 'main']);
+    await git(fixture.cloneB, ['fetch', 'origin']);
+    await git(fixture.cloneB, ['merge', '--ff-only', 'origin/main']);
+    const pulled = await captureJson(() =>
+      teamSyncPull(fixture.cloneB, {
+        task: formattedTaskRef,
+        json: true,
+      }),
+    );
+    expect(pulled).toMatchObject({
+      exitCode: 0,
+      value: {
+        remoteRevision: 3,
+        materializedBundles: [{ status: 'unchanged' }],
+      },
+    });
+    const resumed = await captureJson<CommandJson>(() =>
+      contextResume(fixture.cloneB, formattedTaskRef, {
+        session: SESSION_B,
+        client: 'vitest-b',
+        json: true,
+      }),
+    );
+    expect(resumed).toMatchObject({
+      exitCode: 0,
+      value: { taskRef, taskRevision: 1 },
+    });
+
+    const unconfirmedClaim = await captureJson<CommandJson>(() =>
+      teamClaim(fixture.cloneB, {
+        task: formattedTaskRef,
+        expectedTaskRevision: '1',
+        paths: ['src/**'],
+        session: SESSION_B,
+        client: 'vitest-b',
+        sync: true,
+        json: true,
+      }),
+    );
+    expect(unconfirmedClaim).toMatchObject({
+      exitCode: 3,
+      value: {
+        error: { code: 'MANCODE_MANTEAM_PLAN_CONFIRMATION_REQUIRED' },
+      },
+    });
+    await expect(
+      new V3ContextStore(fixture.cloneB).readTaskSnapshot(taskRef),
+    ).resolves.toMatchObject({
+      metadata: { revision: 1, ownershipEpoch: 0 },
+    });
+
+    const checkpoint = await createV3Checkpoint({
+      projectRoot: fixture.cloneA,
+      taskRef,
+      sessionId: SESSION_A,
+      expectedTaskRevision: 1,
+      kind: 'handoff_offered',
+      summary: 'Publish a journaled handoff checkpoint after initial sync.',
+      nextAction: 'Pull the successor bundle in the second clone.',
+      checkpointId: id(17),
+      operationId: id(18),
+      now: at(3),
+    });
+    expect(
+      await gitOutput(fixture.cloneA, [
+        'status',
+        '--porcelain=v1',
+        '--untracked-files=all',
+      ]),
+    ).toContain(taskRoot);
+    const updatedSync = await captureJson<CommandJson>(() =>
+      teamSyncPush(fixture.cloneA, {
+        task: formattedTaskRef,
+        expectedTaskRevision: String(checkpoint.metadata.revision),
+        session: SESSION_A,
+        client: 'vitest-a',
+        json: true,
+      }),
+    );
+    expect(updatedSync).toMatchObject({
+      exitCode: 0,
+      value: {
+        remoteRevision: 4,
+        taskRevision: checkpoint.metadata.revision,
+        ownershipEpoch: 0,
+      },
+    });
+    const updatedPull = await captureJson(() =>
+      teamSyncPull(fixture.cloneB, { task: formattedTaskRef, json: true }),
+    );
+    expect(updatedPull).toMatchObject({
+      exitCode: 0,
+      value: {
+        remoteRevision: 4,
+        materializedBundles: [
+          {
+            status: 'updated',
+            taskRevision: checkpoint.metadata.revision,
+          },
+        ],
+      },
+    });
+  }, 20_000);
+
   it('fences claims, materializes a quarantined bundle, and transfers ownership once', async () => {
     const fixture = await createFixture();
     const created = await createRemoteTask(fixture.cloneA, fixture.codeHead);
@@ -119,7 +355,7 @@ describe('git-ref coordination across independent clones', () => {
     });
     expect(claim).toMatchObject({
       remoteRevision: 2,
-      ownershipEpoch: 1,
+      ownershipEpoch: 0,
       claim: { state: 'active', ownerActorId: ACTOR_A },
     });
 
@@ -267,7 +503,7 @@ describe('git-ref coordination across independent clones', () => {
         actorId: ACTOR_A,
         taskRef: created.taskRef,
         expectedRemoteRevision: beforeAcceptManifest.revision,
-        expectedOwnershipEpoch: 1,
+        expectedOwnershipEpoch: 0,
         claimId: claim.claim.claimId,
         expectedClaimRevision: 1,
         expiresAt: at(2 * 24 * 60).toISOString(),
@@ -289,11 +525,11 @@ describe('git-ref coordination across independent clones', () => {
     expect(preparedRemoteRevision).toBe(5);
     expect(accepted).toMatchObject({
       remoteRevision: 5,
-      ownershipEpoch: 2,
+      ownershipEpoch: 1,
       handoff: { state: 'accepted', revision: 3 },
       taskBundle: {
         taskRevision: initialBundle.taskRevision + 2,
-        ownershipEpoch: 2,
+        ownershipEpoch: 1,
       },
       forwardRepair: { remoteRevision: 5, ownerActorId: ACTOR_B },
     });
@@ -303,13 +539,13 @@ describe('git-ref coordination across independent clones', () => {
       manifest: {
         revision: 5,
         ownershipFences: [
-          { ownerActorId: ACTOR_B, ownershipEpoch: 2, remoteRevision: 5 },
+          { ownerActorId: ACTOR_B, ownershipEpoch: 1, remoteRevision: 5 },
         ],
         receipts: expect.arrayContaining([
           expect.objectContaining({
             operationId: id(31),
             remoteRevision: 5,
-            ownershipEpoch: 2,
+            ownershipEpoch: 1,
           }),
         ]),
       },
@@ -319,7 +555,7 @@ describe('git-ref coordination across independent clones', () => {
     ).resolves.toMatchObject({
       metadata: {
         ownerActorId: ACTOR_B,
-        ownershipEpoch: 2,
+        ownershipEpoch: 1,
         revision: initialBundle.taskRevision + 2,
         transitionState: 'stable',
       },
@@ -460,7 +696,7 @@ describe('git-ref coordination across independent clones', () => {
           { taskRevision: blocked.metadata.revision, lastOperationId: id(45) },
         ],
         taskBundles: [
-          { taskRevision: blocked.metadata.revision, ownershipEpoch: 1 },
+          { taskRevision: blocked.metadata.revision, ownershipEpoch: 0 },
         ],
       },
     });
@@ -922,9 +1158,9 @@ async function setGitRefConfig(projectRoot: string): Promise<void> {
   );
 }
 
-async function captureJson(
+async function captureJson<T extends Record<string, unknown> = SyncPullJson>(
   action: () => Promise<number>,
-): Promise<{ exitCode: number; value: SyncPullJson }> {
+): Promise<{ exitCode: number; value: T }> {
   const writes: string[] = [];
   const previous = console.log;
   console.log = (value: unknown) => writes.push(String(value));
@@ -932,7 +1168,7 @@ async function captureJson(
     const exitCode = await action();
     return {
       exitCode,
-      value: JSON.parse(writes.at(-1) ?? '{}') as SyncPullJson,
+      value: JSON.parse(writes.at(-1) ?? '{}') as T,
     };
   } finally {
     console.log = previous;
@@ -945,6 +1181,15 @@ interface SyncPullJson extends Record<string, unknown> {
   >;
 }
 
+interface CommandJson extends Record<string, unknown> {
+  error?: { code: string };
+}
+
+interface WorkflowCreateJson extends CommandJson {
+  taskRef: { namespace: 'shared'; taskId: Ulid };
+  metadata: { revision: number; ownershipEpoch: number };
+}
+
 interface WorkflowCompletionJson extends Record<string, unknown> {
   metadata: {
     revision: number;
diff --git a/tests/status.test.ts b/tests/status.test.ts
index 81cf0eb..ce53417 100644
--- a/tests/status.test.ts
+++ b/tests/status.test.ts
@@ -5,6 +5,7 @@ import { afterEach, beforeEach, describe, expect, it } from 'vitest';
 import { init } from '../src/commands/init.js';
 import { install } from '../src/commands/install.js';
 import {
+  type ContinuityStatusResult,
   EXIT_CORRUPT_STATE,
   EXIT_NOT_INITIALIZED,
   EXIT_OK,
@@ -175,12 +176,40 @@ describe('mancode status', () => {
     const logs = await captureLog(() => status(dir));
     const output = logs.join('\n');
 
+    expect(output).toContain(`mancode Continuity v${VERSION}`);
     expect(output).toContain('Activation:  active');
     expect(output).toContain('mancode adapter status:');
     expect(output).not.toContain('V3');
     expect(output).not.toContain('v3_active');
   });
 
+  it('--brief --json exposes only the public Continuity runtime view', async () => {
+    await silentInit(dir, { v3: true, platform: 'codex' });
+
+    const logs = await captureLog(() =>
+      status(dir, { brief: true, json: true }),
+    );
+    const output = logs.join('\n');
+    const result: ContinuityStatusResult = JSON.parse(output);
+
+    expect(result).toMatchObject({
+      schemaVersion: 1,
+      runtime: 'mancode-continuity',
+      state: 'active',
+      version: VERSION,
+      project: path.basename(dir),
+      ready: true,
+      identity: { actorId: null, displayName: null },
+      session: null,
+      task: null,
+      blockers: [],
+    });
+    expect(result).not.toHaveProperty('authority');
+    expect(result).not.toHaveProperty('activation');
+    expect(output).not.toContain('"v3"');
+    expect(output).not.toContain('v3_active');
+  });
+
   it('omits top-level Hooks section for non-Claude Code projects', async () => {
     await silentInit(dir, { platform: 'codex' });
 
diff --git a/tests/v3-adapter-contracts.test.ts b/tests/v3-adapter-contracts.test.ts
index 0cbc025..d7aa4a9 100644
--- a/tests/v3-adapter-contracts.test.ts
+++ b/tests/v3-adapter-contracts.test.ts
@@ -116,13 +116,16 @@ describe('V3 adapter bootstrap integration', () => {
       expect(bootstrap).toContain('# mancode bootstrap');
       expect(bootstrap).toContain('mancode context show --purpose orient');
       expect(bootstrap).toContain('--session <id>');
-      expect(bootstrap).toContain('First run `mancode status --json`');
+      expect(bootstrap).toContain('mancode status --brief --json');
+      expect(bootstrap).toContain(
+        'snapshot already obtained in this conversation',
+      );
       expect(bootstrap).toContain(
         'In operator-facing narration, say `mancode`',
       );
       expect(bootstrap).toContain('mancode team identity create --name');
       expect(bootstrap).toContain(
-        '`currentTask: null` and `MANCODE_TASK_REQUIRED` do not make a session stale',
+        '`task: null` and `MANCODE_TASK_REQUIRED` do not make a session stale',
       );
       expect(bootstrap).toContain(
         'Do not probe workflow subcommands to work around `MANCODE_TASK_REQUIRED`',
@@ -175,7 +178,10 @@ describe('V3 adapter bootstrap integration', () => {
         expect(entry).toContain('## Enter through mancode');
         expect(entry).toContain('In operator-facing narration, say `mancode`');
         expect(entry).not.toMatch(/\bV3\b/);
-        expect(entry).toContain('mancode status --json');
+        expect(entry).toContain('mancode status --brief --json');
+        expect(entry).toContain(
+          'snapshot already obtained in this conversation',
+        );
         expect(entry).not.toContain('.mancode/state.json');
         if (mode !== 'manps') {
           const sessionCommands = Array.from(
diff --git a/tests/version.test.ts b/tests/version.test.ts
index a2e1ae4..05b95de 100644
--- a/tests/version.test.ts
+++ b/tests/version.test.ts
@@ -21,8 +21,8 @@ describe('version', () => {
 
     for (const readme of readmes) {
       expect(readme).toContain(`v${VERSION}`);
-      expect(readme).toContain(`status-V3%20beta%20v${VERSION}`);
-      expect(readme).not.toContain(`status-stable%20v${VERSION}`);
+      expect(readme).toContain(`status-Continuity%20v${VERSION}`);
+      expect(readme).not.toMatch(/V3%20beta/i);
     }
   });
 });
diff --git a/tests/website-docs.test.ts b/tests/website-docs.test.ts
index 964d95f..686ca72 100644
--- a/tests/website-docs.test.ts
+++ b/tests/website-docs.test.ts
@@ -9,6 +9,13 @@ async function readPage(name: string): Promise<string> {
   return readFile(path.join(website, name), 'utf8');
 }
 
+async function readPackageVersion(): Promise<string> {
+  const packageJson = JSON.parse(
+    await readFile(path.join(root, 'package.json'), 'utf8'),
+  ) as { version: string };
+  return packageJson.version;
+}
+
 function internalAnchors(html: string): string[] {
   return [...html.matchAll(/href="#([^"]+)"/g)].map((match) => match[1]);
 }
@@ -71,13 +78,13 @@ describe('website documentation', () => {
     );
   });
 
-  it('shows the V3 workflow command contract', async () => {
+  it('shows the Continuity workflow command contract', async () => {
     for (const name of ['docs.html', 'docs.zh-CN.html']) {
       const html = await readPage(name);
       const start = html.indexOf(
         name === 'docs.html'
-          ? '<h3>A valid V3 sequence</h3>'
-          : '<h3>一条合法的 V3 顺序</h3>',
+          ? '<h3>A valid Continuity sequence</h3>'
+          : '<h3>一条合法的 Continuity 顺序</h3>',
       );
       const end = html.indexOf(
         name === 'docs.html'
@@ -145,8 +152,9 @@ describe('website documentation', () => {
     expect(chinese).toContain('代码，避免');
     expect(chinese).toContain('AI 屎山。');
     expect(chinese).toContain('预览适配器');
-    expect(english).toContain('V3 Beta / v0.3.14');
-    expect(chinese).toContain('V3 Beta / v0.3.14');
+    const version = await readPackageVersion();
+    expect(english).toContain(`Continuity / v${version}`);
+    expect(chinese).toContain(`Continuity / v${version}`);
   });
 
   it('documents the cross-session boundary in both languages', async () => {
@@ -160,16 +168,14 @@ describe('website documentation', () => {
   });
 
   it('keeps version labels aligned with package.json', async () => {
-    const packageJson = JSON.parse(
-      await readFile(path.join(root, 'package.json'), 'utf8'),
-    ) as { version: string };
+    const version = await readPackageVersion();
     for (const name of [
       'index.html',
       'index.zh-CN.html',
       'docs.html',
       'docs.zh-CN.html',
     ]) {
-      expect(await readPage(name)).toContain(`v${packageJson.version}`);
+      expect(await readPage(name)).toContain(`v${version}`);
     }
   });
 });
diff --git a/website/docs.html b/website/docs.html
index 605447b..5b32e8d 100644
--- a/website/docs.html
+++ b/website/docs.html
@@ -24,7 +24,7 @@
 
     <header class="docs-topbar">
       <a class="brand" href="./index.html" aria-label="mancode home"><img src="./logo.png" alt="" /><span>mancode</span><i aria-hidden="true"></i></a>
-      <span class="docs-wordmark">Documentation / v0.3.14</span>
+      <span class="docs-wordmark">Documentation / v0.3.16</span>
       <div class="docs-top-links">
         <a href="./index.html">Website</a>
         <a href="https://github.com/whitelonng/mancode">GitHub</a>
@@ -90,7 +90,7 @@ <h3>Non-interactive or targeted initialization</h3>
           <h3>Verify the installation</h3>
           <div class="code-wrap"><pre><code><span class="command-prompt">$</span> mancode status
 <span class="command-prompt">$</span> mancode status --json</code></pre><button type="button" data-copy-code>Copy</button></div>
-          <p>The human-readable result names the project, current mode, team state, installed adapters, and each adapter's <strong>ready</strong> or <strong>not ready</strong> status. JSON is the stable choice for scripts. A ready adapter means its required generated files are present; it does not guarantee that an already-open agent session has reloaded them.</p>
+          <p>The human-readable result names the project, current mode, team state, installed adapters, and each adapter's <strong>ready</strong> or <strong>not ready</strong> status. JSON is the stable choice for scripts; <code>mancode status --brief --json</code> returns the compact public Continuity view used by coding agents. A ready adapter means its required generated files are present; it does not guarantee that an already-open agent session has reloaded them.</p>
           <div class="docs-admonition"><strong>Reload after setup</strong><p>Restart or reload the coding agent after initialization. Then start a small request in <code>solo</code>. If an adapter remains <em>not ready</em>, repair it with <code>mancode install &lt;platform&gt; --force</code>.</p></div>
         </section>
 
@@ -98,9 +98,9 @@ <h3>Verify the installation</h3>
           <h2>Files &amp; privacy</h2>
           <p>mancode works locally, scans only the current project, and sends no telemetry. It deliberately does not rewrite your project's <code>.gitignore</code>, so your team—not the installer—decides which generated context belongs in version control.</p>
           <table class="docs-table artifact-table"><thead><tr><th>Path</th><th>What it contains</th><th>Before committing</th></tr></thead><tbody>
-            <tr><td><code>.mancode/schema.json</code><br /><code>.mancode/shared/</code><br /><code>.mancode/local/</code></td><td>V3 authority, project policy, sessions, workflow metadata, and local runtime records. Legacy <code>state.json</code> exists only after explicit <code>--legacy</code> initialization.</td><td>Review workflow evidence and local records before sharing; raw host session keys are never persisted.</td></tr>
+            <tr><td><code>.mancode/schema.json</code><br /><code>.mancode/shared/</code><br /><code>.mancode/local/</code></td><td>Continuity authority, project policy, sessions, workflow metadata, and local runtime records. Legacy <code>state.json</code> exists only after explicit <code>--legacy</code> initialization.</td><td>Review workflow evidence and local records before sharing; raw host session keys are never persisted.</td></tr>
             <tr><td><code>.mancode/shared/context/project.json</code><br /><code>.mancode/local/cache/style-tokens.json</code></td><td>Detected stack, validation commands, and local UI tokens.</td><td>Inspect for internal project details and decide whether reproducibility justifies sharing.</td></tr>
-            <tr><td><code>.mancode/local/workflows/</code><br /><code>.mancode/shared/workflows/</code></td><td>Requirements, plans, review and verification ledgers, reports, checkpoints, and summaries.</td><td>Review carefully. Evidence and reports may contain sensitive data; shared workflows pass the V3 privacy boundary.</td></tr>
+            <tr><td><code>.mancode/local/workflows/</code><br /><code>.mancode/shared/workflows/</code></td><td>Requirements, plans, review and verification ledgers, reports, checkpoints, and summaries.</td><td>Review carefully. Evidence and reports may contain sensitive data; shared workflows pass the Continuity privacy boundary.</td></tr>
             <tr><td><code>.mancode/shared/context/decisions/</code><br /><code>.mancode/shared/team/</code></td><td>Confirmed decisions and durable coordination authority.</td><td>Commit only information intentionally written for the team.</td></tr>
             <tr><td><code>.claude/</code>, <code>.cursor/</code>, <code>.agents/</code>, <code>.github/</code>, <code>AGENTS.md</code></td><td>Adapter instructions, commands or prompts, skills, hooks, and managed blocks.</td><td>Share when the adapter is part of the team's development workflow. Content outside mancode's managed blocks remains yours.</td></tr>
           </tbody></table>
@@ -111,7 +111,7 @@ <h2>Files &amp; privacy</h2>
           <h2>Continue work across sessions</h2>
           <p>mancode stores goals, requirements, plans, validation results, and handoff notes under a stable <code>TaskRef</code>. A new chat window or supported CLI can resume the same task and load a purpose-specific Context Pack.</p>
           <div class="docs-note"><strong>What is carried forward</strong><p>This is task continuity, not a copy of raw chat history. Sessions from different clients remain isolated; the new session must identify its own client and explicitly resume the existing TaskRef.</p></div>
-          <div class="code-wrap"><pre><code><span class="command-prompt">$</span> mancode status --json
+          <div class="code-wrap"><pre><code><span class="command-prompt">$</span> mancode status --brief --json
 <span class="command-prompt">$</span> mancode context session new --client claude-code
 <span class="command-prompt">$</span> mancode context resume &lt;namespace:ULID&gt; --session &lt;id&gt; --client claude-code
 <span class="command-prompt">$</span> mancode context show --purpose orient --session &lt;id&gt; --client claude-code</code></pre><button type="button" data-copy-code>Copy</button></div>
@@ -169,21 +169,21 @@ <h2>CLI reference</h2>
           <div class="command-list">
             <article class="command-card"><header><h3>mancode init</h3><span>Start</span></header><p>Creates <code>.mancode/</code>, detects the project, scans style tokens where relevant, and installs adapters.</p><ul><li>Options: <code>--force</code>, <code>--yes</code>, <code>--team</code>, <code>--no-team</code>, <code>--style &lt;name&gt;</code>, <code>--platform &lt;list&gt;</code>, <code>--empty</code>, <code>--lang &lt;locale&gt;</code>.</li></ul></article>
             <article class="command-card"><header><h3>mancode install [platform]</h3><span>Adapter</span></header><p>Installs or repairs an adapter after initialization. Omit the platform for interactive selection.</p><ul><li><code>--force</code> regenerates managed content.</li><li><code>--minimal</code> installs solo-mode essentials only.</li></ul></article>
-            <article class="command-card"><header><h3>mancode status [--json]</h3><span>Inspect</span></header><p>Reports project facts, current mode and workflow, adapter readiness, and Claude hook registration where applicable.</p></article>
+            <article class="command-card"><header><h3>mancode status [--json] [--brief]</h3><span>Inspect</span></header><p>Reports project facts, current mode and workflow, adapter readiness, and Claude hook registration where applicable. Combine <code>--brief --json</code> for the compact Continuity runtime view.</p></article>
             <article class="command-card"><header><h3>mancode list-platforms</h3><span>Discover</span></header><p>Lists adapters known to the installed CLI and marks those already configured in the project.</p></article>
-            <article class="command-card"><header><h3>mancode workflow &lt;subcommand&gt;</h3><span>Govern</span></header><p>Creates and validates V3 requirements, plans, verification evidence, reviews, remediation, and completion.</p><ul><li>Inspect with <code>list</code> and <code>show &lt;namespace:ULID&gt; [--json]</code>.</li><li>Use <code>context compact --dry-run</code> to inspect removable runtime records; V3 workflow authority is not deleted by <code>workflow clean</code>.</li></ul></article>
+            <article class="command-card"><header><h3>mancode workflow &lt;subcommand&gt;</h3><span>Govern</span></header><p>Creates and validates Continuity requirements, plans, verification evidence, reviews, remediation, and completion.</p><ul><li>Inspect with <code>list</code> and <code>show &lt;namespace:ULID&gt; [--json]</code>.</li><li>Use <code>context compact --dry-run</code> to inspect removable runtime records; Continuity workflow authority is not deleted by <code>workflow clean</code>.</li></ul></article>
             <article class="command-card"><header><h3>mancode manps [area]</h3><span>Scan</span></header><p>Runs a deterministic health scan for <code>all</code>, <code>deps</code>, <code>security</code>, <code>dead-code</code>, <code>config</code>.</p><ul><li><code>--json</code> emits machine-readable output.</li><li><code>--remediate</code> enters the explicit remediation path; the default is scan-only.</li></ul></article>
             <article class="command-card"><header><h3>mancode refresh-project</h3><span>Rescan</span></header><p>Refreshes project facts after adding Git, a manifest, a framework, or validation commands, then updates installed static adapters.</p></article>
             <article class="command-card"><header><h3>mancode refresh-style</h3><span>Design</span></header><p>Refreshes the project profile and design tokens. Reinstall static adapters with <code>--force</code> afterward.</p></article>
-            <article class="command-card"><header><h3>mancode uninstall [platform]</h3><span>Remove</span></header><p>Removes one V3 adapter. V3 protects authority from bulk removal; use <code>context compact --dry-run</code> for retention candidates. The <code>--all</code> form is legacy-only.</p></article>
+            <article class="command-card"><header><h3>mancode uninstall [platform]</h3><span>Remove</span></header><p>Removes one Continuity adapter. Continuity protects authority from bulk removal; use <code>context compact --dry-run</code> for retention candidates. The <code>--all</code> form is legacy-only.</p></article>
             <article class="command-card"><header><h3>mancode version</h3><span>Version</span></header><p>Prints the installed CLI version for upgrade checks and issue reports.</p></article>
           </div>
         </section>
 
         <section class="doc-section" id="workflow" data-searchable>
           <h2>Study the /man workflow</h2>
-          <p><code>/man</code> is a progressive nine-step protocol. It first makes the decision inspectable, then lets the user choose between a lightweight handoff, a plan-only result, or governed execution. Every gate is backed by V3 files under <code>.mancode/&lt;namespace&gt;/workflows/&lt;ULID&gt;/</code>.</p>
-          <div class="docs-note"><strong>Who runs these commands?</strong><p>The installed V3 mode entry creates the task and reads its Context Pack. The commands below are the machine contract for studying, debugging, or integrating the process. V3 identifies tasks with explicit <code>namespace:ULID</code> TaskRefs and does not use a legacy active pointer.</p></div>
+          <p><code>/man</code> is a progressive nine-step protocol. It first makes the decision inspectable, then lets the user choose between a lightweight handoff, a plan-only result, or governed execution. Every gate is backed by Continuity files under <code>.mancode/&lt;namespace&gt;/workflows/&lt;ULID&gt;/</code>.</p>
+          <div class="docs-note"><strong>Who runs these commands?</strong><p>The installed Continuity mode entry creates the task and reads its Context Pack. The commands below are the machine contract for studying, debugging, or integrating the process. Continuity identifies tasks with explicit <code>namespace:ULID</code> TaskRefs and does not use a legacy active pointer.</p></div>
 
           <h3>The nine steps and their gates</h3>
           <table class="docs-table workflow-table"><thead><tr><th>Step</th><th>Agent work</th><th>Durable gate</th></tr></thead><tbody>
@@ -239,8 +239,8 @@ <h3>The Step 4 decision</h3>
           </div>
           <p>All three choices require an active <code>man</code> workflow, confirmed requirements, and <code>plan.md</code>. Revise the plan with <code>workflow plan ... revise</code>; every successful mutation returns the revision required by the next command.</p>
 
-          <h3>A valid V3 sequence</h3>
-          <p>V3 uses explicit TaskRefs and expected revisions. The CLI owns every durable mutation; do not edit metadata files or use the legacy <code>--step</code> protocol.</p>
+          <h3>A valid Continuity sequence</h3>
+          <p>Continuity uses explicit TaskRefs and expected revisions. The CLI owns every durable mutation; do not edit metadata files or use the legacy <code>--step</code> protocol.</p>
           <div class="code-wrap"><pre><code><span class="command-prompt"># Bootstrap an actor and explicit session</span>
 <span class="command-prompt">$</span> mancode team identity create --name "Your name"
 <span class="command-prompt">$</span> mancode context session new --client codex --json
@@ -262,7 +262,7 @@ <h3>A valid V3 sequence</h3>
 
           <h3>Verification, manual checks, and remediation</h3>
           <table class="docs-table"><thead><tr><th>Case</th><th>Required record</th><th>Gate behavior</th></tr></thead><tbody>
-            <tr><td>Automated</td><td>The canonical verification ledger records the criterion, command, exit code, evidence summary, and result.</td><td><code>workflow verify ... apply</code> rejects a ledger that does not satisfy the V3 schema and current revision.</td></tr>
+            <tr><td>Automated</td><td>The canonical verification ledger records the criterion, command, exit code, evidence summary, and result.</td><td><code>workflow verify ... apply</code> rejects a ledger that does not satisfy the Continuity schema and current revision.</td></tr>
             <tr><td>Manual</td><td>The ledger records why automation cannot decide and only changes to passed after explicit user confirmation.</td><td>Completion remains blocked while required manual evidence is pending.</td></tr>
             <tr><td>Review finding</td><td>The canonical review ledger stores stable blocker IDs and their remediation state.</td><td>Applying a review ledger makes earlier verification stale; apply current verification after review and before completion.</td></tr>
             <tr><td>User skips review</td><td>The review ledger records the explicit user decision and residual risk.</td><td>Completion still requires a valid summary and every remaining gate.</td></tr>
@@ -279,8 +279,8 @@ <h3>Workflow subcommands</h3>
             <tr><td><code>plan</code></td><td>Revise a plan or confirm its execution decision.</td><td><code>revise --file</code>; <code>confirm --plan-decision</code>; both require revision and session.</td></tr>
             <tr><td><code>update</code></td><td>Change lifecycle status or a blocking reason.</td><td><code>--status --expected-revision --session</code>; governed fields use dedicated commands.</td></tr>
             <tr><td><code>handoff</code></td><td>Transfer a confirmed local plan to solo and later close it.</td><td><code>--to solo</code> or <code>--complete</code>, with revision and session.</td></tr>
-            <tr><td><code>verify / review</code></td><td>Apply canonical V3 ledgers.</td><td><code>apply --file --expected-revision --session</code>.</td></tr>
-            <tr><td><code>list / show / clean</code></td><td>List or inspect V3 TaskRefs. <code>clean</code> is not a V3 authority deletion command.</td><td><code>show &lt;namespace:ULID&gt; --json</code>; use <code>context compact --dry-run</code> for retention.</td></tr>
+            <tr><td><code>verify / review</code></td><td>Apply canonical Continuity ledgers.</td><td><code>apply --file --expected-revision --session</code>.</td></tr>
+            <tr><td><code>list / show / clean</code></td><td>List or inspect Continuity TaskRefs. <code>clean</code> is not a Continuity authority deletion command.</td><td><code>show &lt;namespace:ULID&gt; --json</code>; use <code>context compact --dry-run</code> for retention.</td></tr>
           </tbody></table>
         </section>
 
@@ -307,7 +307,7 @@ <h2>Refresh project context</h2>
 
         <section class="doc-section" id="uninstall" data-searchable>
           <h2>Uninstall safely</h2>
-          <p>Uninstalling one adapter preserves user-authored rules, instructions, unrelated Claude settings, and V3 workflow authority. Use <code>context compact --dry-run</code> to inspect eligible runtime retention records.</p>
+          <p>Uninstalling one adapter preserves user-authored rules, instructions, unrelated Claude settings, and Continuity workflow authority. Use <code>context compact --dry-run</code> to inspect eligible runtime retention records.</p>
           <h3>Remove one adapter</h3>
           <div class="code-wrap"><pre><code><span class="command-prompt">$</span> mancode uninstall cursor --force</code></pre><button type="button" data-copy-code>Copy</button></div>
           <h3>Reinstall one adapter</h3>
@@ -328,7 +328,7 @@ <h3>Claude Code hooks do not run</h3>
           <h3>Cursor rules do not trigger</h3>
           <p>Confirm that <code>.cursor/rules/mancode-*.mdc</code> and <code>.cursor/commands/*.md</code> exist. Core context and solo rules are persistent; higher-intensity rules are description-triggered, so invoke the generated mode command explicitly.</p>
           <h3>A workflow command is rejected</h3>
-          <p>Treat the rejection as a gate, not a file-corruption problem. Use <code>mancode workflow list --json</code> to discover TaskRefs, then <code>mancode workflow show &lt;namespace:ULID&gt; --json</code> to inspect metadata, revision, aggregate, and blockers. Do not edit V3 metadata manually.</p>
+          <p>Treat the rejection as a gate, not a file-corruption problem. Use <code>mancode workflow list --json</code> to discover TaskRefs, then <code>mancode workflow show &lt;namespace:ULID&gt; --json</code> to inspect metadata, revision, aggregate, and blockers. Do not edit Continuity metadata manually.</p>
           <h3>Project facts became stale</h3>
           <p>After adding Git, a manifest, dependencies, or validation scripts, run <code>mancode refresh-project</code>. Use <code>refresh-style</code> for UI tokens. Then repair a static adapter only if its generated files need regeneration.</p>
         </section>
diff --git a/website/docs.zh-CN.html b/website/docs.zh-CN.html
index 0ca73be..02f2a7c 100644
--- a/website/docs.zh-CN.html
+++ b/website/docs.zh-CN.html
@@ -24,7 +24,7 @@
 
     <header class="docs-topbar">
       <a class="brand" href="./index.zh-CN.html" aria-label="mancode 首页"><img src="./logo.png" alt="" /><span>mancode</span><i aria-hidden="true"></i></a>
-      <span class="docs-wordmark">文档 / v0.3.14</span>
+      <span class="docs-wordmark">文档 / v0.3.16</span>
       <div class="docs-top-links">
         <a href="./index.zh-CN.html">官网</a>
         <a href="https://github.com/whitelonng/mancode">GitHub</a>
@@ -90,7 +90,7 @@ <h3>指定平台或非交互初始化</h3>
           <h3>验证安装结果</h3>
           <div class="code-wrap"><pre><code><span class="command-prompt">$</span> mancode status
 <span class="command-prompt">$</span> mancode status --json</code></pre><button type="button" data-copy-code>复制</button></div>
-          <p>普通输出会列出项目、当前模式、团队状态、已安装平台，以及每个平台的 <strong>ready</strong> 或 <strong>not ready</strong> 状态；脚本应使用 JSON。<em>ready</em> 表示适配器要求的生成文件已经存在，但不代表一个早已打开的 Agent 会话已经重新加载这些文件。</p>
+          <p>普通输出会列出项目、当前模式、团队状态、已安装平台，以及每个平台的 <strong>ready</strong> 或 <strong>not ready</strong> 状态；脚本应使用 JSON，编码 Agent 可用 <code>mancode status --brief --json</code> 获取精简的 Continuity 公开视图。<em>ready</em> 表示适配器要求的生成文件已经存在，但不代表一个早已打开的 Agent 会话已经重新加载这些文件。</p>
           <div class="docs-admonition"><strong>初始化后重载</strong><p>安装后重启或重新加载编程 Agent，再用 <code>solo</code> 完成一个小任务。如果平台仍显示 <em>not ready</em>，执行 <code>mancode install &lt;platform&gt; --force</code> 修复。</p></div>
         </section>
 
@@ -98,9 +98,9 @@ <h3>验证安装结果</h3>
           <h2>文件与隐私</h2>
           <p>mancode 本地运行，只扫描当前项目，不发送遥测。它不会擅自改写项目的 <code>.gitignore</code>，因此哪些生成上下文应该进入版本控制，需要由你的团队明确决定。</p>
           <table class="docs-table artifact-table"><thead><tr><th>路径</th><th>包含什么</th><th>提交前怎么处理</th></tr></thead><tbody>
-            <tr><td><code>.mancode/schema.json</code><br /><code>.mancode/shared/</code><br /><code>.mancode/local/</code></td><td>V3 权威、项目策略、会话、工作流元数据和本地运行时记录。只有显式使用 <code>--legacy</code> 初始化后才会出现旧 <code>state.json</code>。</td><td>共享前检查工作流证据和本地记录；原始宿主 session key 不会持久化。</td></tr>
+            <tr><td><code>.mancode/schema.json</code><br /><code>.mancode/shared/</code><br /><code>.mancode/local/</code></td><td>Continuity 权威、项目策略、会话、工作流元数据和本地运行时记录。只有显式使用 <code>--legacy</code> 初始化后才会出现旧 <code>state.json</code>。</td><td>共享前检查工作流证据和本地记录；原始宿主 session key 不会持久化。</td></tr>
             <tr><td><code>.mancode/shared/context/project.json</code><br /><code>.mancode/local/cache/style-tokens.json</code></td><td>识别出的技术栈、验证命令和本地 UI token。</td><td>检查是否带有内部项目信息，再权衡复现能力与暴露范围。</td></tr>
-            <tr><td><code>.mancode/local/workflows/</code><br /><code>.mancode/shared/workflows/</code></td><td>需求、计划、评审和验证台账、报告、checkpoint 与总结。</td><td>重点检查。证据和报告可能含敏感数据；shared workflow 必须通过 V3 隐私边界。</td></tr>
+            <tr><td><code>.mancode/local/workflows/</code><br /><code>.mancode/shared/workflows/</code></td><td>需求、计划、评审和验证台账、报告、checkpoint 与总结。</td><td>重点检查。证据和报告可能含敏感数据；shared workflow 必须通过 Continuity 隐私边界。</td></tr>
             <tr><td><code>.mancode/shared/context/decisions/</code><br /><code>.mancode/shared/team/</code></td><td>已确认决策与持久化协调权威。</td><td>只提交有意写给团队的内容。</td></tr>
             <tr><td><code>.claude/</code>、<code>.cursor/</code>、<code>.agents/</code>、<code>.github/</code>、<code>AGENTS.md</code></td><td>平台 instructions、命令或 prompt、skills、hooks 与受控区块。</td><td>团队确实把该适配器作为开发流程的一部分时再共享；mancode 管理标记之外的内容仍归用户所有。</td></tr>
           </tbody></table>
@@ -111,7 +111,7 @@ <h2>文件与隐私</h2>
           <h2>跨会话继续工作</h2>
           <p>mancode 把目标、需求、计划、检查结果和交接信息保存在稳定的 <code>TaskRef</code> 下。新聊天窗口或受支持的 CLI 可以恢复同一项任务，再按当前目的读取精简的 Context Pack。</p>
           <div class="docs-note"><strong>会带过去什么</strong><p>这是任务上下文的续接，不是原始聊天记录的复制。不同客户端的 session 仍然隔离；新 session 必须使用自己的 client 身份，并显式恢复已有 TaskRef。</p></div>
-          <div class="code-wrap"><pre><code><span class="command-prompt">$</span> mancode status --json
+          <div class="code-wrap"><pre><code><span class="command-prompt">$</span> mancode status --brief --json
 <span class="command-prompt">$</span> mancode context session new --client claude-code
 <span class="command-prompt">$</span> mancode context resume &lt;namespace:ULID&gt; --session &lt;id&gt; --client claude-code
 <span class="command-prompt">$</span> mancode context show --purpose orient --session &lt;id&gt; --client claude-code</code></pre><button type="button" data-copy-code>复制</button></div>
@@ -169,21 +169,21 @@ <h2>CLI 参考</h2>
           <div class="command-list">
             <article class="command-card"><header><h3>mancode init</h3><span>初始化</span></header><p>创建 <code>.mancode/</code>，识别项目，在适用时扫描样式 token，并安装平台适配器。</p><ul><li>选项：<code>--force</code>、<code>--yes</code>、<code>--team</code>、<code>--no-team</code>、<code>--style &lt;name&gt;</code>、<code>--platform &lt;list&gt;</code>、<code>--empty</code>、<code>--lang &lt;locale&gt;</code>。</li></ul></article>
             <article class="command-card"><header><h3>mancode install [platform]</h3><span>适配器</span></header><p>初始化后安装或修复一个平台。省略平台名时进入交互选择。</p><ul><li><code>--force</code> 重新生成受控内容。</li><li><code>--minimal</code> 只安装 solo 的必要内容。</li></ul></article>
-            <article class="command-card"><header><h3>mancode status [--json]</h3><span>检查</span></header><p>报告项目事实、当前模式和工作流、平台就绪状态，以及适用时的 Claude hook 注册情况。</p></article>
+            <article class="command-card"><header><h3>mancode status [--json] [--brief]</h3><span>检查</span></header><p>报告项目事实、当前模式和工作流、平台就绪状态，以及适用时的 Claude hook 注册情况。组合使用 <code>--brief --json</code> 可获得精简的 Continuity 运行时视图。</p></article>
             <article class="command-card"><header><h3>mancode list-platforms</h3><span>发现</span></header><p>列出当前 CLI 支持的平台，并标出项目中已经配置的平台。</p></article>
-            <article class="command-card"><header><h3>mancode workflow &lt;subcommand&gt;</h3><span>治理</span></header><p>创建和校验 V3 需求、计划、验证证据、评审、修复与完成状态。</p><ul><li>使用 <code>list</code> 和 <code>show &lt;namespace:ULID&gt; [--json]</code> 检查。</li><li>使用 <code>context compact --dry-run</code> 检查可回收的运行时记录；V3 workflow authority 不会被 <code>workflow clean</code> 删除。</li></ul></article>
+            <article class="command-card"><header><h3>mancode workflow &lt;subcommand&gt;</h3><span>治理</span></header><p>创建和校验 Continuity 需求、计划、验证证据、评审、修复与完成状态。</p><ul><li>使用 <code>list</code> 和 <code>show &lt;namespace:ULID&gt; [--json]</code> 检查。</li><li>使用 <code>context compact --dry-run</code> 检查可回收的运行时记录；Continuity workflow authority 不会被 <code>workflow clean</code> 删除。</li></ul></article>
             <article class="command-card"><header><h3>mancode manps [area]</h3><span>扫描</span></header><p>对 <code>all</code>、<code>deps</code>、<code>security</code>、<code>dead-code</code> 或 <code>config</code> 执行确定性健康扫描。</p><ul><li><code>--json</code> 输出机器可读结果。</li><li><code>--remediate</code> 显式进入修复路径；默认只扫描。</li></ul></article>
             <article class="command-card"><header><h3>mancode refresh-project</h3><span>重扫项目</span></header><p>在加入 Git、manifest、框架或验证命令后刷新项目事实，并更新已安装的静态适配器。</p></article>
             <article class="command-card"><header><h3>mancode refresh-style</h3><span>设计上下文</span></header><p>刷新项目画像和设计 token，之后可用 <code>--force</code> 重装需要更新的静态适配器。</p></article>
-            <article class="command-card"><header><h3>mancode uninstall [platform]</h3><span>移除</span></header><p>移除一个 V3 适配器。V3 会保护权威数据不被批量删除；用 <code>context compact --dry-run</code> 检查保留候选。<code>--all</code> 只适用于 legacy 项目。</p></article>
+            <article class="command-card"><header><h3>mancode uninstall [platform]</h3><span>移除</span></header><p>移除一个 Continuity 适配器。Continuity 会保护权威数据不被批量删除；用 <code>context compact --dry-run</code> 检查保留候选。<code>--all</code> 只适用于 legacy 项目。</p></article>
             <article class="command-card"><header><h3>mancode version</h3><span>版本</span></header><p>输出已安装 CLI 版本，便于升级检查和提交问题。</p></article>
           </div>
         </section>
 
         <section class="doc-section" id="workflow" data-searchable>
           <h2>深入研究 /man 工作流</h2>
-          <p><code>/man</code> 是一个渐进式九步协议：先让决策变得可检查，再由用户选择轻量交接、仅保留计划，或继续治理式执行。每个闸门都由 V3 的 <code>.mancode/&lt;namespace&gt;/workflows/&lt;ULID&gt;/</code> 文件支撑。</p>
-          <div class="docs-note"><strong>谁来运行这些命令？</strong><p>安装后的 V3 mode entry 会创建任务并读取 Context Pack。下面展示的是供研究、排错或集成使用的机器契约。V3 使用显式 <code>namespace:ULID</code> TaskRef，不依赖旧的活动指针。</p></div>
+          <p><code>/man</code> 是一个渐进式九步协议：先让决策变得可检查，再由用户选择轻量交接、仅保留计划，或继续治理式执行。每个闸门都由 Continuity 的 <code>.mancode/&lt;namespace&gt;/workflows/&lt;ULID&gt;/</code> 文件支撑。</p>
+          <div class="docs-note"><strong>谁来运行这些命令？</strong><p>安装后的 Continuity mode entry 会创建任务并读取 Context Pack。下面展示的是供研究、排错或集成使用的机器契约。Continuity 使用显式 <code>namespace:ULID</code> TaskRef，不依赖旧的活动指针。</p></div>
 
           <h3>九个步骤与闸门</h3>
           <table class="docs-table workflow-table"><thead><tr><th>步骤</th><th>Agent 工作</th><th>持久化闸门</th></tr></thead><tbody>
@@ -239,8 +239,8 @@ <h3>Step 4 的三种决策</h3>
           </div>
           <p>三种选择都要求活动中的 <code>man</code> 工作流、已确认需求和 <code>plan.md</code>。使用 <code>workflow plan ... revise</code> 修订计划；每次成功写入都会返回下一个命令需要的 revision。</p>
 
-          <h3>一条合法的 V3 顺序</h3>
-          <p>V3 使用显式 TaskRef 和 expected revision。所有持久化变更都由 CLI 完成；不要手工编辑 metadata，也不要使用 legacy <code>--step</code> 协议。</p>
+          <h3>一条合法的 Continuity 顺序</h3>
+          <p>Continuity 使用显式 TaskRef 和 expected revision。所有持久化变更都由 CLI 完成；不要手工编辑 metadata，也不要使用 legacy <code>--step</code> 协议。</p>
           <div class="code-wrap"><pre><code><span class="command-prompt"># 创建 actor 和显式 session</span>
 <span class="command-prompt">$</span> mancode team identity create --name "Your name"
 <span class="command-prompt">$</span> mancode context session new --client codex --json
@@ -262,7 +262,7 @@ <h3>一条合法的 V3 顺序</h3>
 
           <h3>自动验证、手工确认与修复后重验</h3>
           <table class="docs-table"><thead><tr><th>场景</th><th>必须记录</th><th>闸门行为</th></tr></thead><tbody>
-            <tr><td>自动验证</td><td>规范化 verification ledger 记录验收项、命令、退出码、证据摘要与结果。</td><td><code>workflow verify ... apply</code> 会拒绝不符合 V3 schema 或 revision 的台账。</td></tr>
+            <tr><td>自动验证</td><td>规范化 verification ledger 记录验收项、命令、退出码、证据摘要与结果。</td><td><code>workflow verify ... apply</code> 会拒绝不符合 Continuity schema 或 revision 的台账。</td></tr>
             <tr><td>手工验证</td><td>台账记录自动化不能判断的原因；只有用户明确确认后才能改为 passed。</td><td>必需人工证据未完成时，completion 持续阻塞。</td></tr>
             <tr><td>评审发现</td><td>规范化 review ledger 保存稳定 blocker ID 及其 remediation 状态。</td><td>应用 review ledger 会使旧验证失效；评审后、完成前必须 apply 当前验证。</td></tr>
             <tr><td>用户跳过评审</td><td>review ledger 记录用户显式决定和残余风险。</td><td>仍然需要有效 summary 和其余全部门禁。</td></tr>
@@ -279,8 +279,8 @@ <h3>workflow 子命令速查</h3>
             <tr><td><code>plan</code></td><td>修订计划或确认执行决策。</td><td><code>revise --file</code>；<code>confirm --plan-decision</code>；两者都需要 revision 和 session。</td></tr>
             <tr><td><code>update</code></td><td>改变生命周期状态或阻塞原因。</td><td><code>--status --expected-revision --session</code>；治理字段使用专用命令。</td></tr>
             <tr><td><code>handoff</code></td><td>把已确认的本地计划交给 solo，随后关闭。</td><td><code>--to solo</code> 或 <code>--complete</code>，并传 revision 与 session。</td></tr>
-            <tr><td><code>verify / review</code></td><td>应用规范化 V3 台账。</td><td><code>apply --file --expected-revision --session</code>。</td></tr>
-            <tr><td><code>list / show / clean</code></td><td>列出或检查 V3 TaskRef；<code>clean</code> 不是删除 V3 权威数据的命令。</td><td><code>show &lt;namespace:ULID&gt; --json</code>；保留清理使用 <code>context compact --dry-run</code>。</td></tr>
+            <tr><td><code>verify / review</code></td><td>应用规范化 Continuity 台账。</td><td><code>apply --file --expected-revision --session</code>。</td></tr>
+            <tr><td><code>list / show / clean</code></td><td>列出或检查 Continuity TaskRef；<code>clean</code> 不是删除 Continuity 权威数据的命令。</td><td><code>show &lt;namespace:ULID&gt; --json</code>；保留清理使用 <code>context compact --dry-run</code>。</td></tr>
           </tbody></table>
         </section>
 
@@ -307,7 +307,7 @@ <h2>刷新项目上下文</h2>
 
         <section class="doc-section" id="uninstall" data-searchable>
           <h2>安全卸载</h2>
-          <p>卸载单个平台会保留用户自己编写的 rules、instructions、无关的 Claude settings 和 V3 工作流权威数据。使用 <code>context compact --dry-run</code> 检查可清理的运行时保留记录。</p>
+          <p>卸载单个平台会保留用户自己编写的 rules、instructions、无关的 Claude settings 和 Continuity 工作流权威数据。使用 <code>context compact --dry-run</code> 检查可清理的运行时保留记录。</p>
           <h3>移除一个适配器</h3>
           <div class="code-wrap"><pre><code><span class="command-prompt">$</span> mancode uninstall cursor --force</code></pre><button type="button" data-copy-code>复制</button></div>
           <h3>重装一个适配器</h3>
@@ -328,7 +328,7 @@ <h3>Claude Code hooks 没有触发</h3>
           <h3>Cursor rules 没有触发</h3>
           <p>确认 <code>.cursor/rules/mancode-*.mdc</code> 和 <code>.cursor/commands/*.md</code> 都存在。核心上下文与 solo rules 持续生效；高强度规则按描述触发，因此应显式调用生成的模式命令。</p>
           <h3>workflow 命令被拒绝</h3>
-          <p>把拒绝当作流程闸门，而不是元数据损坏。先用 <code>mancode workflow list --json</code> 找到 TaskRef，再用 <code>mancode workflow show &lt;namespace:ULID&gt; --json</code> 检查元数据、revision、aggregate 和 blocker。不要手工编辑 V3 metadata。</p>
+          <p>把拒绝当作流程闸门，而不是元数据损坏。先用 <code>mancode workflow list --json</code> 找到 TaskRef，再用 <code>mancode workflow show &lt;namespace:ULID&gt; --json</code> 检查元数据、revision、aggregate 和 blocker。不要手工编辑 Continuity metadata。</p>
           <h3>项目事实已经过期</h3>
           <p>添加 Git、manifest、依赖或验证脚本后执行 <code>mancode refresh-project</code>；UI token 改变时执行 <code>refresh-style</code>。只有静态适配器的生成文件需要更新时，才重新安装对应平台。</p>
         </section>
diff --git a/website/index.html b/website/index.html
index f0528b1..d1a1eeb 100644
--- a/website/index.html
+++ b/website/index.html
@@ -439,7 +439,7 @@ <h2 class="reveal">Stop babysitting<br />AI bloat.</h2>
         <a href="https://github.com/whitelonng/mancode/blob/main/README.md">中文</a>
         <a href="https://github.com/whitelonng/mancode/blob/main/LICENSE">AGPL-3.0</a>
       </div>
-      <span class="footer-status"><i></i> V3 Beta / v0.3.14</span>
+      <span class="footer-status"><i></i> Continuity / v0.3.16</span>
     </footer>
   </body>
 </html>
diff --git a/website/index.zh-CN.html b/website/index.zh-CN.html
index 37cad6e..9c3d691 100644
--- a/website/index.zh-CN.html
+++ b/website/index.zh-CN.html
@@ -88,6 +88,6 @@ <h2>一项任务。<br />下个会话继续。</h2>
 
       <section class="final-cta"><div class="final-cta-mark" aria-hidden="true">M</div><p class="reveal">简单任务少一点仪式。<br />关键任务多一点纪律。</p><h2 class="reveal">别再替 AI<br />收拾臃肿。</h2><div class="final-actions reveal"><div class="command-bar command-light"><code><span>&gt;</span> npm install -g mancode</code><button type="button" data-copy="npm install -g mancode">复制</button></div><a class="button button-dark" href="https://github.com/whitelonng/mancode">GitHub 上查看 ↗</a></div></section>
     </main>
-    <footer class="site-footer"><a class="brand footer-brand" href="#top"><img src="./logo.png" alt="" /><span>mancode</span></a><p>AI 编码 Agent 工作流调度工具。<br />日常训练，关键任务季后赛。</p><div class="footer-links"><a href="https://github.com/whitelonng/mancode">GitHub</a><a href="./docs.zh-CN.html">中文文档</a><a href="./docs.html">English</a><a href="https://github.com/whitelonng/mancode/blob/main/LICENSE">AGPL-3.0</a></div><span class="footer-status"><i></i> V3 Beta / v0.3.14</span></footer>
+    <footer class="site-footer"><a class="brand footer-brand" href="#top"><img src="./logo.png" alt="" /><span>mancode</span></a><p>AI 编码 Agent 工作流调度工具。<br />日常训练，关键任务季后赛。</p><div class="footer-links"><a href="https://github.com/whitelonng/mancode">GitHub</a><a href="./docs.zh-CN.html">中文文档</a><a href="./docs.html">English</a><a href="https://github.com/whitelonng/mancode/blob/main/LICENSE">AGPL-3.0</a></div><span class="footer-status"><i></i> Continuity / v0.3.16</span></footer>
   </body>
 </html>