Show what you made with Open Film Skills

[62656456]

Use this thread to share a public or anonymized result made with Open Film Skills.

Please include:

• Skill used
• What you asked it to do
• What it produced
• What worked
• The one change that would improve it most
• Agent host and model, if relevant

You can link an image, video, document, or the smallest output excerpt that lets others judge the result. A structural pass is not the same as a good creative result, so honest failures are welcome too.

Before posting, remove private scripts, client or company information, credentials, unpublished assets, faces you cannot share, and personal contact details.

New to the repository? Start with the verified Storyboard Director 5.4.4 example: https://github.com/62656456/ai-film-skills/blob/main/examples/storyboard-director-5.4.4-visible-camera-plan.md

[62656456]

Maintainer example: Director Agent diagnosis and revision

https://github.com/62656456/ai-film-skills/blob/main/examples/director-agent-before-after.md

A deliberately weak scene begins with an unclear objective, arbitrary refusal, symmetrical dialogue, and no state change. The revised version adds a concrete deadline, visible alternatives, supporting-character self-protection, money/key/object leverage, and a consequential exit state.

Evidence status: author-run internal forward test, SELF-AUDIT ONLY. No independent cold reader, external-user adoption, creative stability, or user acceptance is claimed.

Use the same format when sharing a result: request, output, what worked, what failed, and the one change that would improve it most.

[62656456]

Maintainer media example: playable local 3D previs evidence

https://62656456.github.io/ai-film-skills/#media

The gallery contains:

• a 5.0-second, 960×540 camera-and-blocking previs for push, actor blocking, a hard cut, lateral tracking, and scene parallax;
• a 2.792-second, 960×402 rigged contact gate for two-hand grip, final-surface contact, hold, and opposite recoil.

Each file has a preserved poster, SHA-256, dimensions, duration, status, rights basis, exclusions, and “proves / does not prove” fields in the media manifest.

Evidence status: original local project media authorized for this showcase. These are bounded technical validation clips, not finished AI films or external-user adoption. User-rejected, failed, UI-only, and rights-unclear local assets were excluded.

[62656456]

Maintainer visual-language gallery: one visible frame per genre Skill

https://github.com/62656456/ai-film-skills#explore-the-visual-language

The repository homepage now shows nine directly visible examples:

• eight deployed genre Skills: cyberpunk, epic, fantasy, horror, noir, romance, war, and wuxia;
• one separately labeled experimental hard-science-fiction example.

Selection evidence:

• complete generated-image archive inventory at the freeze: 2,819 PNG files, 6.49 GB, 208 session directories, all readable, 2,819 unique SHA-256 values, zero duplicates;
• seven images newly generated from the current Skill contracts;
• two prior original atlas panels reused only after provenance, earlier QA, and isolated-crop review;
• nine final files, nine unique hashes, explicit visible contracts and user-acceptance boundaries.

The archive inventory is not presented as 2,819 individual aesthetic approvals. The gallery illustrates observable design parameters and maintainer internal selection; it does not prove that a Skill alone generated an image, cross-model stability, external adoption, production readiness, or user aesthetic approval.

Manifest: https://github.com/62656456/ai-film-skills/blob/main/docs/style-gallery/manifest.json