lightflow-video-auto-edit-plan

# LightFlow Video Auto Edit Plan Use `lightflow.video_auto_edit_plan` only after `lightflow.video_highlights` has returned its approved VideoScore `clips` output. Pass that output unchanged as `clips`. The package-owned Rust runner produces a structured `lightflow.video.edit-plan.v1` from those explicit, HMAC-verified ranges; `lightflow.video_render_edit` consumes the result. ## Runtime The workflow crate contains a Rust-native `runner.v1` binary, so installation does not depend on repository-relative scripts. It does not perform ASR, semantic selection, silence removal, or scene detection. ## Contract - Workflow id: `lightflow.video_auto_edit_plan` - Required input `clips`: JSON array of explicit clip objects with VideoScore provenance. - Required input `brief`: human editing goal, story outline, or script notes. - Optional input `style`: editing style guidance. Defaults to `clean social edit`. - Optional input `constraints`: JSON delivery constraints for aspect ratio, duration, frame rate, width, and height. - Output `edit_plan`: JSON edit decision plan. - Output `summary`: readable planning summary. ## CLI Usage ```bash # Set this only in the runtime environment. Use the same key for # lightflow.video_highlights and lightflow.video_auto_edit_plan. export LIGHTFLOW_VIDEOSCORE_EVIDENCE_KEY='at-least-32-non-whitespace-bytes' lfw run lightflow.video_auto_edit_plan \ --input clips='[{"id":"intro","path":"intro.mp4","start":0,"end":8,"highlight":{"workflow":"lightflow.video_highlights","source_path":"intro.mp4","start_seconds":0,"end_seconds":8,"score":3.7,"model":"TIGER-Lab/VideoScore-v1.1","reason":"Clear opening vehicle shot.","evidence":"<generated-by-lightflow.video_highlights>"}}]' \ --input brief='"Cut a concise launch recap with a strong opening hook."' \ --input style='"fast social product edit"' \ --input constraints='{"aspect_ratio":"9:16","max_duration_seconds":45}' ``` `<generated-by-lightflow.video_highlights>` is a schema placeholder, not a usable value. Copy the actual lowercase-hex HMAC-SHA256 tag produced by `lightflow.video_highlights`; do not forge it. ## API Usage Start `lfw serve`, then call the workflow through the shared HTTP run contract: ```bash curl -sS -X POST http://127.0.0.1:5174/workflows/lightflow.video_auto_edit_plan/run \ -H 'content-type: application/json' \ -d '{"inputs":{"clips":[{"id":"intro","path":"intro.mp4","start":0,"end":8,"highlight":{"workflow":"lightflow.video_highlights","source_path":"intro.mp4","start_seconds":0,"end_seconds":8,"score":3.7,"model":"TIGER-Lab/VideoScore-v1.1","reason":"Clear opening vehicle shot.","evidence":"<generated-by-lightflow.video_highlights>"}}],"brief":"Cut a concise launch recap with a strong opening hook.","style":"fast social product edit","constraints":{"aspect_ratio":"9:16","max_duration_seconds":45}}}' ``` ## Change Notes When changing inputs, outputs, runtime behavior, or common commands, update this skill and the runner tests in the same change so agents can safely run and inspect the workflow.