v2.1.0 — Remotion, HyperFrames and Openverse in production; PDF items and quieter tabs in review

Make, review and publish social content from an AI assistant: five skills that work together, with the hosted ViewPrinter MCP server.

Assets

File	For
viewprinter-skill-2.1.0.zip	All five skills combined into one skill, with SKILL.md at the archive root. This is what npx skills add https://viewprinter.tech fetches.
viewprinter-2.1.0.zip	The full plugin for Claude Code and Codex: manifests, five skills and MCP config.
viewprinter-social-manager-2.1.0.zip	The ClawHub package: the same combined skill, in a folder.

All three can be rebuilt from this tag byte for byte: timestamps are fixed, file modes come from the path, and the build host isn't recorded.

7f5b6fda9efed44ba790c01a1718b6df7a456f79fa725d0dc575f2f628c0b2bc  viewprinter-2.1.0.zip
82cb48eaf8524b0dfaec0f887230f285fd3160e090ec71233b0a1c05ba391d1a  viewprinter-social-manager-2.1.0.zip
a227c3e0ba2e0b2ba9e33e9d133fe73438fd3574e36e6e7158785b29935c530d  viewprinter-skill-2.1.0.zip

This release

• New production tools. The tools check and guide now cover:
  • Remotion for 3D device shots: a lit device with the real recording as its screen, landing on the edit's first frame;
  • HyperFrames for HTML motion graphics, explainers, slideshows and templated batches filled at render time;
  • Openverse for CC sound effects, ambience and music beds, with the license and attribution recorded in the archive.
    Footage editing still defaults to Tesseract.
• The review workspace:
  • shows a script, storyboard or brief as one PDF item, with open and download links;
  • releases a hidden tab's live connection and pauses its media;
  • reuses unchanged summaries instead of rebuilding them on every check;
  • can scope save, release, drop and clean-sources to one batch with --batch.
• Publishing: moving weaker posts later when stronger content would exceed a posting cap, planning a reflow without writing, moving a creative to another account without leaving a duplicate, upload purposes after checksum deduplication, and links to the exact dashboard post.
• Production and learning guidance: generation modes, voice casting, narration with sound off, scene coherence, export quality and frame-exact cuts; post age at capture time, comment intent and research filters.
• The archive accepts the <project>/<id> form that find prints, wherever it takes an ID.

The local helpers need Python 3.9 or newer and run on macOS, Linux and Windows. CI runs every skill's tests on all three, with Python 3.9 and 3.12.

This release doesn't change the hosted service or refresh installed client caches. The directory listings are separate distribution channels.