Lumos-789 · Nov 5, 2026
@@ -0,0 +1,33 @@
+{
+  "name": "delivery-video",
+  "version": "1.1.2",
+  "description": "Turn a feature that actually runs into a trustworthy delivery demo video, or a defect into a traceable feedback video: real screen recordings, edge-tts narration, word-synced subtitles and a deterministic quality gate.",
+  "author": {
+    "name": "Lumos-789",
+    "url": "https://github.com/Lumos-789"
+  },
+  "homepage": "https://github.com/Lumos-789/delivery-video-plugin",
+  "repository": "https://github.com/Lumos-789/delivery-video-plugin",
+  "license": "MIT",
+  "keywords": [
+    "delivery",
+    "demo-video",
+    "bug-report",
+    "screen-recording",
+    "quality-gate",
+    "pi",
+    "dsh"
+  ],
+  "skills": "./skills/",
+  "interface": {
+    "displayName": "Delivery Video",
+    "shortDescription": "Record and gate delivery demo videos",
+    "longDescription": "Records real screen sessions of working software, renders narration and word-synced subtitles, and runs a deterministic quality gate before the video is delivered. The same pipeline produces defect/feedback videos with severity and repro/actual chips plus a hashed issue report. No static screenshots, mocks or stale footage.",
+    "developerName": "Lumos-789",
+    "category": "Productivity",
+    "capabilities": [
+      "Read",
+      "Write with confirmation"
+    ]
+  }
+}
@@ -0,0 +1,27 @@
+{
+  "name": "delivery-video",
+  "version": "1.1.2",
+  "description": "Record, render and gate trustworthy delivery demo videos — or defect/feedback report videos — of real, working software: real screen recordings, edge-tts narration, word-synced subtitles and a deterministic quality gate.",
+  "description_i18n": {
+    "en": "Record, render and gate trustworthy delivery demo videos — or defect/feedback report videos — of real, working software: real screen recordings, edge-tts narration, word-synced subtitles and a deterministic quality gate.",
+    "zh-CN": "把真实运行的功能录成可交付的演示视频，或把缺陷录成可追溯的问题反馈视频：真实录屏、edge-tts 旁白、逐字对齐字幕与确定性质量闸，产出 1920x1080 成片与完整证据链。"
+  },
+  "author": {
+    "name": "Lumos-789",
+    "url": "https://github.com/Lumos-789"
+  },
+  "homepage": "https://github.com/Lumos-789/delivery-video-plugin",
+  "repository": "https://github.com/Lumos-789/delivery-video-plugin",
+  "license": "MIT",
+  "keywords": [
+    "delivery",
+    "demo-video",
+    "bug-report",
+    "screen-recording",
+    "acceptance",
+    "quality-gate",
+    "edge-tts",
+    "ffmpeg"
+  ],
+  "skills": "skills"
+}
@@ -0,0 +1,30 @@
+# Changelog
+
+## 1.1.2
+
+- Feedback reports are now fully bound to the evidence chain. `render.py` writes the
+  artifact SHA256 plus, into `quality_gate/report.json`, a normalized hash of the
+  `feedback.json` fields and the exact `feedback.generated.md` bytes.
+- `verify_delivery.py` recomputes both hashes. Editing `issue.title`/`issue.actual` (or
+  any other report field), editing the markdown, deleting the markdown, or rendering
+  with an older report that has no feedback binding now fails with a specific message
+  instead of `DELIVERY=PASS`. The video-side binding (artifact SHA256, truncated video,
+  stripped audio, re-encoded resolution) is unchanged.
+- Tests: `tests/test_feedback_binding.py` covers clean pass, edited title, edited actual,
+  edited markdown and deleted markdown; `tests/test_render.py` covers canonical hash
+  normalization. Feedback artifacts rendered by 1.1.1 must be re-rendered to gain the
+  binding.
+
+## 1.1.1
+
+- Feedback/defect report mode: `intent: "feedback"` plus an `issue` block adds severity
+  and phase chips, a repro-steps overlay, and writes `feedback.json` and
+  `feedback.generated.md` bound to the artifact SHA256. Quality-gate rules are unchanged.
+
+## 1.0.x (2026-10-31 → 2026-11-02)
+
+- 1.0.0 — first ZCode release: real-recording pipeline (`preflight`/`render`/`verify`), the built-in `delivery-v1` gate, bilingual READMEs.
+- 1.0.1 — narration/TTS, subtitle line-breaking and low-bitrate false-positive fixes; preflight checks render dependencies early.
+- 1.0.2 — default warnings cleaned up; auditable notices kept.
+- 1.0.3 / 1.0.4 — DSH runtime skill registration, then registration scoped to each Agent (plugin shell: `package.json` + `index.js` + `cordis.patch.yml`).
+- 1.0.5 / 1.0.6 — desktop-environment hardening; case-driven recording guidance.
@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Lumos-789
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
@@ -0,0 +1,76 @@
+# delivery-video
+
+A skill plugin that turns a feature that **actually runs** into a trustworthy delivery demo video — real screen recordings, synthesized narration, word-synced subtitles, and a deterministic quality gate that refuses to pass a fake or broken demo.
+
+Most "AI demo videos" are stitched from static screenshots, mock pages or stale recordings. This plugin enforces the opposite discipline: every segment must point at a real recording with a traceable time range, the last shot must end on a verified result state, and the artifact must pass hard checks for resolution/FPS/codec, audio presence, loudness, black frames and freeze frames before anything is posted. Bitrate density is reported as an informational notice; text still needs a 1:1 visual check. The uploaded file is always the gated file, bound by SHA256.
+
+In DeepSeek Harness, the skill is registered in both the Host and each live Agent scope. That lets the installed plugin win over an older same-name copy in `~/.agents/skills`, even when an Agent preset mounts the filesystem skill provider in a nearer layer.
+
+## What it does
+
+1. **Verify first, record second** — the workflow requires evidence that the current version really runs (git SHA, tickets, request IDs) before any recording.
+2. **Case-driven structure (recommended)** — for multi-case deliveries the skill suggests listing the acceptance cases first, recording one real run per case, then assembling segments in execution order with a result shot per case. Optional, not a gate.
+3. **episode.json recipe** — each segment binds one real take + time range + one narration line + optional crop/case chip.
+   Set `intent: "feedback"` with an `issue` block to produce a defect report instead of a
+   delivery: severity and phase chips, a repro-steps overlay, and both `feedback.json` and a
+   ticket-ready `feedback.generated.md`. The gate report binds the artifact SHA256, a
+   normalized hash of the `feedback.json` fields and the exact markdown bytes, so
+   `verify_delivery.py` fails when the report was edited after rendering, or the markdown is
+   missing or was edited after rendering. Copy `examples/feedback-episode.example.json` as a
+   starting point. The same quality gate applies.
+4. **Self-contained renderer** — `render.py` synthesizes narration via edge-tts (word-boundary clocks), times each segment to its narration (0.3x–3x speed is hard-checked, never clamped — out of range fails), draws the label chip and word-synced subtitles, concatenates, loudness-normalizes to −14 LUFS, and runs the built-in quality gate. Output: `screen_demo.mp4` (1920×1080, 30fps, h264 with AAC audio — a non-AAC track is a gate warning `E102`, not a failure), `timeline.json`, cover/preview frames, `quality_gate/report.json`.
+5. **Fail-loud checks** — `preflight.py` validates the recipe before rendering (structure, intervals, crops, browser-recording evidence, TTS cache contracts, narration red lines); `verify_delivery.py` verifies the finished artifact, gate binding and optional independent visual review afterwards.
+6. **Attachment gate** — `render.py gate` runs the same hard checks on any mp4 before it goes into a ticket, chat or customer channel.
+
+## Prerequisites
+
+- `python3` 3.9+
+- `ffmpeg` and `ffprobe` on PATH
+- `pip install "edge-tts>=7.3.1" Pillow`
+
+`edge-tts` 7.2.8 and older can return HTTP 403 from the upstream endpoint. Use 7.3.1 or newer; the renderer fails early with an upgrade command when the installed version is too old.
+Run `preflight.py`, `render.py` and `verify_delivery.py` with the same Python interpreter. Preflight fails early when ffmpeg/ffprobe, Pillow, a supported edge-tts version or a CJK font is missing.
+
+Platform setup:
+
+- macOS: `brew install ffmpeg`, then `python3 -m pip install "edge-tts>=7.3.1" Pillow`
+- Windows: install FFmpeg with `winget install --id Gyan.FFmpeg -e` (or `choco install ffmpeg`). If `python3` is not available, use `py -3` consistently and run `py -3 -m pip install "edge-tts>=7.3.1" Pillow`
+- Linux: install `ffmpeg` and a CJK font such as `fonts-noto-cjk` with your package manager, then install the Python packages
+
+The renderer needs a CJK font for Chinese labels and subtitles. macOS and Windows normally provide one; minimal Linux images often do not, and preflight reports the missing font before rendering.
+
+## Side effects (disclosed for review)
+
+- **Network**: narration synthesis calls the public edge-tts (Microsoft Cognitive Services) endpoint; no other network access.
+- **Files**: writes only inside the build directory you pass in (`tts/` cache, per-segment clips, final mp4, gate reports).
+- **Commands**: runs `ffmpeg`/`ffprobe` locally; the skill never uploads, publishes or deletes anything by itself.
+
+## Quick start
+
+Install: `dsh plugin add /path/to/delivery-video-plugin` (DSH bundle). It exposes the skill in `skills/delivery-video/`.
+
+1. Record real takes of the feature with your usual tool (macOS `Cmd+Shift+5`, OBS, or a scripted browser screencast — see the skill's recording reference). For two or more acceptance cases, list them first using `examples/cases.example.md` as a template, then record one real run per case.
+2. Create `build/screen_demo-<slug>/episode.json` (copy `examples/episode.example.json`).
+3. Run the three scripts (use `py -3` instead of `python3` on Windows when needed):
+
+```bash
+python3 <skill>/scripts/preflight.py build/screen_demo-myfeature
+python3 <skill>/scripts/render.py render build/screen_demo-myfeature
+python3 <skill>/scripts/verify_delivery.py build/screen_demo-myfeature
+```
+
+Run the unit tests with:
+
+```bash
+python3 -m unittest discover -s tests -v
+```
+
+Suggested trial prompt after install: *"用 delivery-video 把这个项目刚完成的功能录成一段 90 秒交付演示视频"* — the skill will walk verification, storyboard, recording guidance, recipe assembly, render and gating.
+
+## Same-name skill copies
+
+ZCode 0.17.x can keep a user-level skill and a plugin skill with the same name. If an older `~/.agents/skills/delivery-video` still exists, the unqualified `/skill delivery-video` command may select that copy. Select `delivery-video:delivery-video` in the composer, or use `/skill delivery-video:delivery-video <task>`, to run the installed plugin.
+
+## License
+
+MIT © Lumos-789
@@ -0,0 +1,74 @@
+# delivery-video
+
+把**真实运行**的功能做成可信交付演示视频的 skill 插件——真实录屏、合成旁白、逐字对齐字幕，加一道拒绝放过造假或残次演示的确定性质量闸。
+
+大多数「AI 演示视频」是静态截图、mock 页面或旧录屏拼出来的。这个插件强制相反的纪律：每个镜头必须指向一段真实录屏和可追溯的时间区间，最后一镜必须落在已回读验证的结果状态上，成片必须先过硬检查（分辨率/帧率/编码、音轨、响度、黑帧、冻结帧）才允许外发。码率密度会作为信息项保留，关键页面仍需按 1:1 目检。贴出去的文件永远是过闸的那一个文件，用 SHA256 绑定。
+
+在 DeepSeek Harness 中，该 Skill 会同时注册到 Host 与每个 Agent 作用域。因此即使 Agent preset 在更近的文件系统层加载了 Skill，已安装插件仍会优先于 `~/.agents/skills` 里的旧同名副本。
+
+## 它做什么
+
+1. **先验真，再录**——工作流要求先有当前版本真实跑通的版本证据（git SHA、工单、请求 ID），才允许录制。
+2. **按用例组织（推荐）**——多 case 交付时，建议先拆一张用例清单，再一个 case 一次真实运行录制，最后按执行顺序拼镜头，每个 case 以自己的结果回读收尾。属使用建议，不做强制校验。
+3. **episode.json 配方**——每个镜头绑定一段真实原片 + 时间区间 + 一句旁白 + 可选裁切/case 徽标。
+   写 `intent: "feedback"` 并添加 `issue` 块即可产出问题反馈视频：严重级别与阶段徽标、
+   复现步骤浮层、复现/现象镜头，以及 `feedback.json` 和可直接贴工单的
+   `feedback.generated.md`。闸报告同时绑定成片 SHA256、`feedback.json` 字段的归一化哈希
+   与 markdown 原始字节哈希，渲染后改动报告内容或删除/篡改 markdown 都会被
+   `verify_delivery.py` 拒绝；质量闸规则不变。可从 `examples/feedback-episode.example.json` 复制起步。
+4. **自包含渲染器**——`render.py` 用 edge-tts 合成旁白（逐字时间钟），按旁白时长配速每个镜头（0.3x–3x 变速硬校验，超范围直接失败，绝不自动钳制），绘制标签徽标与逐字同步字幕，拼接、响度归一到 −14 LUFS，并自动跑内置质量闸。输出：`screen_demo.mp4`（1920×1080、30fps、h264，音轨为 AAC——非 AAC 只会得到闸警告 `E102`，不算失败）、`timeline.json`、封面/预览帧、`quality_gate/report.json`。
+5. **Fail-loud 检查**——`preflight.py` 渲染前校验配方（结构、区间、裁切、浏览器录屏证据、TTS 缓存契约、旁白红线）；`verify_delivery.py` 渲染后核验成片、闸绑定与可选的独立视觉复核。
+6. **附件闸**——`render.py gate` 对任意 mp4 跑同一套硬检查，贴进工单、群聊或客户渠道之前必须过。
+
+## 前置依赖
+
+- `python3` 3.9+
+- PATH 里有 `ffmpeg` 和 `ffprobe`
+- `pip install "edge-tts>=7.3.1" Pillow`
+
+`edge-tts` 7.2.8 及更早版本会从上游端点收到 HTTP 403。请使用 7.3.1 或更新版本；版本过低时渲染器会直接失败并给出升级命令。
+`preflight.py`、`render.py` 和 `verify_delivery.py` 要用同一个 Python 解释器运行。预检会在 ffmpeg/ffprobe、Pillow、edge-tts 版本或中文字体不满足要求时直接失败。
+
+不同系统的准备方式：
+
+- macOS：`brew install ffmpeg`，然后执行 `python3 -m pip install "edge-tts>=7.3.1" Pillow`
+- Windows：用 `winget install --id Gyan.FFmpeg -e`（或 `choco install ffmpeg`）安装 FFmpeg；如果没有 `python3`，统一改用 `py -3`，并执行 `py -3 -m pip install "edge-tts>=7.3.1" Pillow`
+- Linux：用包管理器安装 `ffmpeg` 和中文字体（例如 `fonts-noto-cjk`），再安装 Python 包
+
+渲染器需要中文字体来绘制标签和字幕。macOS 和 Windows 通常自带；精简 Linux 镜像可能没有，预检会在渲染前直接报告字体缺失。
+
+## 副作用披露（审核要求如实说明）
+
+- **网络**：旁白合成调用公共 edge-tts（Microsoft 认知服务）端点；无其他网络访问。
+- **文件**：只在你传入的构建目录内写入（`tts/` 缓存、逐镜头片段、成片、闸报告）。
+- **命令**：本地执行 `ffmpeg`/`ffprobe`；本 skill 自身不上传、不发布、不删除任何东西。
+
+## 快速开始
+
+安装：`dsh plugin add /path/to/delivery-video-plugin`（DSH bundle），会暴露 `skills/delivery-video/` 下的 skill。
+
+1. 用你惯用的工具录真实原片（macOS `Cmd+Shift+5`、OBS，或脚本化浏览器录屏——见 skill 内录制指引）。覆盖两条以上验收用例时，先按 `examples/cases.example.md` 拆清单，再一个 case 录一次真实运行。
+2. 创建 `build/screen_demo-<slug>/episode.json`（可复制 `examples/episode.example.json`）。
+3. 依次跑三个脚本（Windows 没有 `python3` 时改用 `py -3`）：
+
+```bash
+python3 <skill>/scripts/preflight.py build/screen_demo-myfeature
+python3 <skill>/scripts/render.py render build/screen_demo-myfeature
+python3 <skill>/scripts/verify_delivery.py build/screen_demo-myfeature
+```
+
+单测命令：
+
+```bash
+python3 -m unittest discover -s tests -v
+```
+
+安装后的试用 prompt：*"用 delivery-video 把这个项目刚完成的功能录成一段 90 秒交付演示视频"*——skill 会带你走完验真、分镜、录制指引、配方组装、渲染与过闸。
+
+## 同名 Skill 副本
+
+ZCode 0.17.x 允许用户级 Skill 和插件 Skill 同名共存。如果本机还保留旧的 `~/.agents/skills/delivery-video`，直接输入 `/skill delivery-video` 可能选中旧副本。请在输入框选择 `delivery-video:delivery-video`，或使用 `/skill delivery-video:delivery-video <任务>` 调用已安装插件。
+
+## 许可
+
+MIT © Lumos-789
@@ -0,0 +1,3 @@
+- insert:
+    - id: delivery-video
+      name: 'delivery-video'
@@ -0,0 +1,17 @@
+# Acceptance cases (optional)
+
+Use this when one video covers two or more acceptance cases. It is a working
+list, not a gate: keep the same case ids in the take filenames and in
+`segments[].case`. Keep the file in the target project, not in the plugin.
+
+Feature: order owner assignment
+Audience: operations staff
+Problem this demo covers: handing an order to another owner without losing the saved state
+
+| Case | Proves | Entry | Precondition / data | Expected observable result | Result |
+| --- | --- | --- | --- | --- | --- |
+| F-01 | main flow works end to end | order detail page | test order `T-1001` | saved owner is shown after refresh | PASS |
+| F-02 | empty selection is rejected | same page | clear the owner field | inline validation appears, nothing is saved | PASS |
+
+Takes: `F-01-entry.mp4`, `F-01-result.mp4`, `F-02-validation.mp4`.
+Each case starts from a known state and is recorded as its own real run.
@@ -0,0 +1,37 @@
+{
+  "voice": "zh-CN-YunxiNeural",
+  "rate": "+5%",
+  "title": "Order task list — delivery demo",
+  "segments": [
+    {
+      "id": "01-entry",
+      "case": "F-01",
+      "label": "入口",
+      "shot": "take-01.mp4",
+      "start": 0.5,
+      "end": 11.5,
+      "crop": [56, 159, 1512, 738],
+      "text": "从订单详情进入任务运营名单，当前负责人和保存入口都在页面上。"
+    },
+    {
+      "id": "02-save",
+      "case": "F-01",
+      "label": "保存",
+      "shot": "take-02.mp4",
+      "start": 0.3,
+      "end": 9.8,
+      "crop": [56, 159, 1512, 738],
+      "text": "修改负责人后点保存，页面提示已更新，没有报错。"
+    },
+    {
+      "id": "03-readback",
+      "case": "F-01",
+      "label": "回读",
+      "shot": "take-03.mp4",
+      "start": 0.4,
+      "end": 8.2,
+      "crop": [56, 159, 1512, 738],
+      "text": "刷新页面重新打开名单，刚才的修改仍然生效。"
+    }
+  ]
+}