v0.1.3

Corrects the AI-disclosure guidance in the README and the nuevo-documental skill: a human voice does not exempt a video from disclosure. Realistic generated scenes must be declared on upload.

v0.1.2

Documentation only: new README with a demo clip from the plugin's own output, the per-video config example, running costs and explicit limits.

v0.1.1

Fixes found by rendering the vertical short through the documented steps:

• The CTA text overlapped the last subtitle during the final 3 s of voice. It now sits in the upper third.
• The audio mix was sized to the voice instead of the video, which cut the CTA tail off the final file.
• A short config without MUSICA/SFX fell back to the demo audio plan and ffmpeg failed. The mix now skips cues past the end of the video and names the missing file; the skill documents the short's own audio plan.

Also: CI now runs the config tests and the dependency check on Linux, macOS and Windows.

v0.1.0

First public release. Three skills — setup, brand-setup, nuevo-documental — and the Python engines behind them.

Run end to end on macOS only; the Linux and Windows install paths in setup are untested. Skills are written in Spanish.