v2.2.1 — Cross-platform release integrity

v2.2.1 — Cross-platform release integrity

v2.2.1 is the recommended patch release for the native-template and visual-quality work introduced in v2.2.0.

Install

pipx install creating-explainer-videos-skill==2.2.1
# or
npm install --global creating-explainer-videos-skill@2.2.1

Fixed

• CI now derives npm tarball and Python wheel filenames from package.json, so release validation cannot silently retain the previous version number.
• The freeze diagnostic requires 2.5 seconds of sustained near-identical frames. This preserves detection of truly stalled output while allowing intentional low-motion explanatory holds across Chromium/FFmpeg platforms.

All compiler, template-native structure, semantic-state, routed-connector, topic-cover, narration coverage, cue-frame, audio normalization, security, and human-release boundaries from v2.2.0 remain unchanged.

Verification

This patch is released only after the full Node/Python test matrix, packed npm and wheel smoke tests, real Linux HyperFrames/Chrome/FFmpeg rendering, and public Registry installation checks pass from the same commit.

v2.2.0 — Native templates and visual quality safeguards

v2.2.0 — Native templates and default visual quality safeguards

v2.2.0 turns the v2.1 topic DSL into an end-to-end visual system. Topic objects, relations, states, cue actions, covers, and QC evidence now flow through the selected template's native structure instead of falling back to a generic card renderer.

Install

pipx install creating-explainer-videos-skill==2.2.0
# or
npm install --global creating-explainer-videos-skill@2.2.0

Then update the installed Skill:

explainer-video-skill install --target codex --json
explainer-video-skill verify --target codex --json

What changed

• The compiler emits Paper Theatre, Spatial Chamber, or Ink Explainer native DOM fingerprints and native motion selectors.
• Spatial Chamber routes use visible, directional, boundary-to-boundary paths and animate every signal route.
• Semantic roles and explicit tones distinguish inputs, controllers, success, warning, failure, and metrics without requiring custom CSS.
• Covers reuse the completed topic scene and therefore carry the same subject objects and relationships as the film.
• Static validation rejects caption-area collisions, unreadably small text containers, missing cue actions, disconnected spatial mechanisms, and undrawn connectors.
• Mechanism IDs and narration mechanismRefs make it possible to verify that the spoken explanation covers the intended mechanism.
• Canonical narration, TTS input, and subtitles must match cue by cue; average cue duration must be at least 2.2 seconds.
• Media QC verifies template structure and topic cover provenance, extracts one representative frame per narration cue, and creates denser contact sheets.
• Final audio is normalized to approximately -16 LUFS during mux.

Forward-validation lesson

A technically decodable MP4 can still be visually generic, omit the selected template's motion language, route lines through text, or use a cover unrelated to the topic. This release converts those observed failure modes into compiler behavior, tests, and release-blocking checks so the intended production method is explicit and repeatable across different Agent environments.

Compatibility

The CLI and Skill name are unchanged. New and rebuilt schema-2 projects must satisfy the stricter mechanism, narration, timing, scene, and QC contracts. Existing published artifacts are untouched. Schema-1 projects retain the v2.0 generic fallback, which remains unsuitable as evidence of a completed topic-specific scene.

See MIGRATION_V2.md for the exact upgrade checklist.

Verification boundary

The release suite covers unit tests, examples, packed npm and wheel installation, real HyperFrames rendering, cover capture, FFmpeg mux, ffprobe inspection, template-structure audit, representative cue frames, and contact sheets. Human review is still required for factual accuracy, teaching clarity, pronunciation, synchronization, and publication approval.

v2.1.0 — Topic visual programs and executable narration

v2.1.0 — Topic visual programs and executable narration

v2.1.0 closes the two largest gaps in the v2 executable scaffold: topic-specific scene construction and runnable narration synthesis.

Install

pipx install creating-explainer-videos-skill==2.1.0
# or
npm install --global creating-explainer-videos-skill@2.1.0

Project and media commands require Node.js 22+, Chrome or Edge, FFmpeg, and ffprobe. The reference real-speech adapter additionally requires a host Python environment with edge-tts installed.

Topic visual program

New projects contain visual-program.json. It describes bounded semantic elements and cue-bound actions, then compiles them into deterministic template-specific DOM/SVG:

explainer-video-skill visual validate .\project --json
explainer-video-skill visual compile .\project --json
explainer-video-skill visual preview .\project --output .\preview.html --json

The DSL supports groups, text, nodes, shapes, connectors, local assets, annotations, normalized geometry, six layouts, and seven deterministic action kinds. Raw HTML, scripts, remote resources, traversal, and unsupported references are rejected.

The three existing templates remain visual skins and motion grammars. They now consume the topic program instead of being limited to the generic input/change/output body. Schema-1 v2.0 projects without a visual program retain the legacy fallback.

Executable narration

python -m pip install edge-tts
explainer-video-skill narration doctor .\project --adapter edge-tts --json
explainer-video-skill narration synthesize .\project --adapter edge-tts --voice zh-CN-YunxiNeural --allow-network --json

The engine synthesizes cue by cue, normalizes to 48 kHz stereo PCM WAV, removes provider silence, measures speech with ffprobe, adds deterministic gaps, writes the narration master, and rebuilds the caption-identical measured cue timeline.

Valid cached cues are reused by input/audio hash and duration. narration recover selectively regenerates only invalid or changed cues.

The hash-trusted host-command protocol allows additional providers without making declarative extensions executable. Network and provider-cost authorization remain explicit, and credentials stay outside project/request/cache files.

fixture-tts exists only for deterministic tests. Its timing evidence is marked test-only and cannot pass the real-audio production gate.

Evidence and limitations

• Credit-card clearing and quantum tunneling now build different topic-object scene programs.
• CI tests the DSL, path/reference validation, adapters, authorization, host protocol, cache recovery, real FFmpeg/ffprobe synthesis, examples, and the existing Chrome/HyperFrames media pipeline.
• The package still does not infer that a generated diagram is factually or pedagogically correct.
• Real Edge TTS requires an explicit network-authorized run and a complete human pronunciation/listening review.
• Automated QC can produce only release_candidate_pending_human_listen; a named human must review the exact candidate hash before passed.

GitHub Release assets contain the npm tarball, portable Skill ZIP, PyPI wheel/source distribution, and SHA256SUMS built from the tagged commit.

v2.0.1 — Release integrity remediation

v2.0.1 — Release-integrity remediation

v2.0.1 makes the public package, documented runtime contract, real renderer, and release evidence agree. It preserves v2.0.0 as historical evidence and supersedes it for new installations.

Install

The dependency-free Python lifecycle CLI works without Node.js:

pipx install creating-explainer-videos-skill==2.0.1
explainer-video-skill --version

Project creation, template preview, rendering, and media commands use the bundled JavaScript runtime and require Node.js 22+, Chrome or Edge, FFmpeg, and ffprobe.

The guaranteed npm-form artifact is the GitHub Release tarball:

npm install --global https://github.com/wzwailr/creating-explainer-videos-skill/releases/download/v2.0.1/creating-explainer-videos-skill-2.0.1.tgz

The npm Registry is an independent publication channel. Check its live state before using it:

npm view creating-explainer-videos-skill version

What changed

• Node.js 22+ now matches the pinned HyperFrames 0.8.15 runtime.
• doctor detects installed or pinned on-demand HyperFrames without downloading during diagnosis.
• Generated renderers satisfy the HyperFrames composition, timeline, font, caption, and safe-geometry contract.
• Cover capture waits for a real non-empty output.
• Short-video contact sheets must exist and be non-empty before the audit can pass.
• Python wheel and source-distribution metadata is pinned to Core Metadata 2.4 for current Twine/registry compatibility.
• CI performs an actual browser frame render, WAV generation, audio/video mux, ffprobe inspection, and automated media audit.

Scope and release boundary

The built-in scene is a generic runnable mechanism scaffold. It does not automatically create subject-specific high-fidelity diagrams, illustrations, or visual direction. The Agent must author those assets and motions from the knowledge structure, then a named human must review the exact candidate hash. Automation can produce only release_candidate_pending_human_listen, never passed.

Release assets include the PyPI wheel/source distribution, npm tarball, portable Skill ZIP, and SHA256SUMS attached to the GitHub Release.

v2.0.0 · 通用可执行解释型视频 Skill

v2.0.0 — Generic executable explainer-video scaffold

v2 turns the project into a general-purpose, installable production system for mechanism, process, system, science, and technical explainer videos.

Highlights

• Generic PyPI/npm distributions, Skill, CLI, repository name, and documentation.
• Executable state-driven Agent loop with evidence gates.
• JSON-first project scaffold with runnable renderer and independent cover.
• Canonical narration/caption identity, underscore speech protection, and measured timing.
• Three official visual templates: Paper Theatre, Spatial Chamber, and Ink Explainer.
• Deterministic render, FFmpeg mux/audit, artifact hashes, publishing package, and explicit human release decision.
• Declarative permission-listed extensions and domain-specific presets.
• Two non-AI fixtures proving the core is not tied to one series.

Install

pipx install creating-explainer-videos-skill==2.0.0
explainer-video-skill --version
npm install --global https://github.com/wzwailr/creating-explainer-videos-skill/releases/download/v2.0.0/creating-explainer-videos-skill-2.0.0.tgz

The Python CLI performs Skill lifecycle operations without Node.js and delegates project/render commands to the canonical bundled runtime with a Node.js 18+ gate.

Release assets include the PyPI wheel/source distribution, npm tarball, portable Skill ZIP, and SHA-256 checksums. v1.1.0 remains available as the legacy AI-series-specific release.

Published artifacts and verification

• PyPI: https://pypi.org/project/creating-explainer-videos-skill/2.0.0/
• GitHub CI: https://github.com/wzwailr/creating-explainer-videos-skill/actions/runs/33353834169
• Public-index smoke test: version, both CLI names, three visual templates, Skill installation, and all 65 manifest files passed in a fresh environment.
• Wheel SHA-256: 2e4f852970f5cd6b33a958201575afcc03222ffc84eed6156a37f09f73caec5e
• Source distribution SHA-256: 0cfe9a049dfb082cd0a6999f152b4b1dcccca40de67b2af2840a0fab6dc16b40

Release boundary

The package can automate source, renderer, and media evidence, but never fabricates human approval. A video is publishable only after a named reviewer fully reviews the exact candidate hash and records passed.

v1.1.0 · Legacy AI-series-specific release

v1.1.0 — 可安装、可扩展的 AI 原理视频生产 Skill

这是首个公开 GitHub Release。它把《AI 底层原理图解》系列已经验证的内容、动画、音画同步、封面和发布质检经验，整理成一个不绑定单一 Agent 产品的可安装 Skill。

主要能力

• 从 GitHub、Release .tgz 或 ZIP 安装。
• 支持 Codex 和任意目录型 Agent Skill 宿主。
• 支持安装、升级、完整性校验、回滚和可恢复卸载。
• 内置 Ink Explainer 与 Paper Theatre 两套视觉系统。
• 视觉、声音、调研、QC、发布五类扩展可独立替换。
• 统一旁白与字幕，以真实 TTS 时轴驱动语义动画。
• 生成独立封面、发布资料和明确的发布决策证据。

Release 附件

• creating-ai-principle-videos-skill-1.1.0.tgz：推荐的 npm/npx 安装包。
• creating-ai-principle-videos.skill.zip：手工安装兼容包。
• SHA256SUMS.txt：两个附件的 SHA-256 校验值。

当前边界

• 尚未发布到 npm Registry；当前推荐直接从 GitHub 或 Release 安装。
• 不包含字体、GSAP 商业插件、音乐、图片和其他第三方媒体二进制文件。
• 自动检查通过不等于视频可发布；完整人工听看和 release_decision: passed 仍是最终门槛。