@seanyao/roll 3.614.3 → 3.615.1

package/CHANGELOG.md

@@ -2,6 +2,30 @@
 
 ## Unreleased
 
+## v3.615.1 — 2026-06-15
+
+### 修复
+
+- **故事档案页的文本证据又能直接看了**:AC 下的 Vitest/test 输出等文本证据改回内联折叠展示,不用点开新页面;截图证据仍是缩略图点开看大图,文件缺失时显示明确空态。(FIX-285) `[dossier]`
+  <!-- evidence: .roll/features/delivery-dossier/FIX-285/latest/FIX-285-report.html -->
+
+- **项目页的选角和钩子信息按正确稿显示了**:Casting 区恢复 3+1 阶梯卡、ramp 竖条和场景角色三列布局;Loop 的 Hooks 区改为列本仓真实 git 钩子,不再把调度 lane 当成 hooks。(FIX-284) `[loop]`
+  <!-- evidence: .roll/features/delivery-dossier/FIX-284/latest/FIX-284-report.html -->
+
+## v3.614.4 — 2026-06-14
+
+### 改进
+
+- **循环实时流对所有 AI 都看得见了**:以前只有 claude 跑的循环能在实时窗里看到在干什么,换成 codex 等别的 AI 就是一片空白。现在用一套统一的"活动信号"把不同 AI 的输出都翻成同一种可读的实时流,还带进度心跳——哪个 AI、跑到哪一步都看得清,长时间没动静也不会让你误以为卡死了。(US-LOOP-077)
+
+### 修复
+
+- **"完成"必须是真合并了**:以前一轮循环可能把没合并进主线、甚至没通过测试的活,用"环境问题"当借口标成已完成。现在没真合并就不算完成,失败的一轮也不再被放过——测试红了就是红了,不拿环境当挡箭牌。(FIX-295)
+- **复杂的活强制独立评审,不许自己评自己**:改动一多就必须有独立评审才放行——优先换一个不同的 AI 来评;如果你本机只装了一个 AI,就用它的另一个独立实例来评,只要不是同一个会话里自己给自己打分。拿不到独立评审就拦下重试,而不是偷偷降级成自评。(FIX-293)
+- **几个 AI 能正常启动了**:agy、gemini 等 AI 之前因为启动参数顺序写反而起不来(连带它们参与评审也起不来),现在修好了。(FIX-296)
+- **showcase 演示按你真实的环境识别 AI**:`roll showcase` 以前在一个空白沙箱里探测你装了哪些 AI,老是误报"没装";现在按你真实环境探测,装了哪些认哪些。(FIX-292)
+- **循环页面只显示真正干过活的循环**:以前调度器每隔半小时"醒来发现没活、又睡下"的空转,也被当成一条循环显示在页面上,几百条把页面刷得没法看;现在只显示真正接了活的循环,循环编号也不再因为截取方式出错显示成负数(像 "-32144")。(FIX-297)
+
 ## v3.614.3 — 2026-06-14
 
 ### 改进

package/README.md

@@ -44,7 +44,8 @@ First time through? Start with [Getting started](guide/en/getting-started.md).
 | Command | Description |
 |---------|-------------|
 | **Autonomy · daily use** | |
-| `roll loop <on\|off\|now\|status\|runs\|log\|story\|events\|eval\|signals\|alert\|fmt\|pr-inbox\|mute\|unmute\|pause\|resume\|reset\|gc>` | Manage the autonomous BACKLOG executor (incl. per-cycle result scoring) |
+| `roll loop <on\|off\|now\|status\|watch\|runs\|log\|story\|events\|eval\|signals\|alert\|fmt\|pr-inbox\|mute\|unmute\|pause\|resume\|reset\|gc>` | Manage the autonomous BACKLOG executor (incl. per-cycle result scoring) |
+| `roll loop watch [-n <lines>] [--verbose\|--raw] [--attach]` | Read-only, concise, real-time view of THIS project's loop — auto-tails `.roll/loop/live.log` through the renderer (cycle / story / results / ALERT / cost / heartbeat). It never writes to or signals a running cycle; Ctrl-C ends the view, not the loop. `--attach` joins the loop's tmux observe window read-only |
 | `roll loop go [--epic <e>\|--cards <ids>] [--budget <usd>] [--for <duration>] [--review <auto\|hetero\|self\|off>]` | Run goal mode manually until the scoped work completes, pauses, or hits a budget/usage/timebox guardrail; it can run while the scheduler is off, and a paused loop should be resumed first; final review defaults to `auto` |
 | `roll loop goal` | Show persisted goal scope, review mode, usage, limits, safety gate, and last decision |
 | `roll brief` | Show latest owner brief |
@@ -61,7 +62,7 @@ First time through? Start with [Getting started](guide/en/getting-started.md).
 | `roll offboard` | Remove Roll from this project |
 | `roll test [--where] [--reset]` | Run the test suite (routes through the isolation adapter; unknown types fail loud) |
 | `roll ci [--wait]` | Show or wait for current commit's CI status |
-| `roll release [--dry-run]` | The ONE release flow: bump → changelog fold → package gate → PR → merge → consistency gate → tag push |
+| `roll release [--dry-run]` | The ONE release flow: bump → changelog fold → package gate → commit-push → consistency gate → PR → auto-merge → tag push (gate runs before merge; drives the merge via GitHub auto-merge) |
 | `roll showcase [--card <ID>]` | The golden-path standard E2E in an isolated sandbox: reset the card, cast a heterogeneous real-agent trio (kimi/claude/pi), deliver via the loop, capture CLI+web screenshots, assemble the evidence chain, emit a pass/fail verdict |
 | `roll pair [init\|status\|score]` | Cross-agent pairing: heterogeneous peer re-checks and scores deliveries |
 | `roll cycles [--since 1d\|3d\|7d\|all]` | The cycle ledger — one line per cycle, failures never swallowed |