[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6BXXCGES8C0HE4HH2SH4SBG":3,"skill-same:01M6BXXCGES8C0HE4HH2SH4SBG":80},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":26,"publishedAt":52,"linkOnly":53,"markdownContent":54,"markdownBody":55,"frontMatter":56,"refName":57,"commitSha":58,"skillDirectory":59,"forksCount":60,"repositoryHtmlUrl":61,"repositoryHomepageUrl":62,"ownerHtmlUrl":63,"ownerBlogUrl":12,"ownerDescription":64,"updatedAt":65,"install":66,"resources":79},"01M6BXXCGES8C0HE4HH2SH4SBG","ad-creative-testing",null,"Design and build matched ad variants that test one storytelling choice at a time (question-first vs direct, a reversal vs the expected outcome, one loop vs several, resolved vs open ending, product facts vs a person's goal), write the test plan, and read the results without mistaking retention for success. Use when someone wants A\u002FB or split-test versions of an ad or marketing video, asks which hook or structure works better, or brings back retention curves, view-through rates or ad results to interpret.","nodetool-ai\u002Fnodetool","nodetool-ai","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F164873569?v=4","01M6BXPFDW36K9K1MR8ZH63HG2","nodetool.ai","01M6BXPFJ893XM219GQN60B0WK",635,"AGPL-3.0","https:\u002F\u002Fgithub.com\u002Fnodetool-ai\u002Fnodetool\u002Fblob\u002Fmain\u002Fpackages\u002Fsystem-skills\u002Fad-creative-testing\u002FSKILL.md",[18,23],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000008","测试与质量","SCENARIO","使用场景",{"categoryNo":24,"categoryName":25,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000016","图像与多媒体",[27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51],"ad","creative","ad creative","testing","ad creative testing","direct","when","variants","which","retention","time","structure","back","question","choice","marketing","plan","loop","open","test plan","Video","first","product","test","design","2026-11-10 10:12:30",false,"---\nname: ad-creative-testing\ndescription: \"Design and build matched ad variants that test one storytelling choice at a time (question-first vs direct, a reversal vs the expected outcome, one loop vs several, resolved vs open ending, product facts vs a person's goal), write the test plan, and read the results without mistaking retention for success. Use when someone wants A\u002FB or split-test versions of an ad or marketing video, asks which hook or structure works better, or brings back retention curves, view-through rates or ad results to interpret.\"\n---\n\n# Testing ad structure with matched variants\n\nCreator frameworks for hooks and story loops have not been tested against\nalternatives in controlled comparisons. The way to know whether a structure\nworks for this product and this audience is to run two versions that differ\nin that structure and nothing else. This skill builds those versions in\nNodeTool, writes the plan the user runs on their ad platform, and reads what\ncomes back.\n\nNodeTool makes and renders the variants. It does not buy media or read ad\naccounts, so the user runs the test and brings the numbers back. Load\n`ad-story-loops` for the loop map the variants are built from, and the job\nskill that made the ad (`motion-ad`, `commercial-beat-sheet`,\n`launch-commercial`, `ugc-video`) for how its board and cut are built.\n\n## One variable per test\n\nA pair of variants answers one question. Hold everything else fixed:\n\n- runtime within 0.5s, the same aspect ratio and the same opening frame\n  unless the opening is the variable\n- the same product assets, cast, music bed, grade and type system\n- the same claims, the same CTA wording and the same end card\n- the same placement, audience, budget and dates on the platform\n\nChange the hook and the ending together and a win cannot be attributed to\neither. A third arm is fine when the user has the budget, but each arm still\ndiffers from the control in one way.\n\n## Variables worth testing\n\n| Test | Control | Variant | Primary outcome |\n|---|---|---|---|\n| Curiosity gap | States the result, then explains | Names the question, explains, then shows the result | Retention through the first payoff |\n| Earned surprise | The expected outcome | A coherent reversal set up earlier | Post-view satisfaction and recall |\n| Loop count | One central question | Several connected questions | Completion and comprehension |\n| Ending | Resolved | Open, pointing at the next episode | Stated intent to watch the next one, and actual continuation |\n| Person vs facts | Product facts alone | The same facts through a person's goal | Brand recall and qualified engagement |\n| Hook archetype | One archetype | Another, same body | 2s and 6s views, then retention |\n\nWrite each test as a hypothesis before building: \"Opening on the question\ninstead of the result keeps more viewers to the first payoff at 8s, without\nlowering click-through.\"\n\n## Build the variants\n\nBuild the control first and finish it. Every variant is a copy with one\nchange, never a second ad made from scratch.\n\n1. **Snapshot the control.** `create_timeline_version {timeline_id}` on the\n   finished cut, so the control can always be restored.\n2. **Copy the cut.** `get_timeline {timeline_id}`, then `create_timeline`\n   with the same `fps`, `width` and `height` and a versioned name\n   (`\u003Cbrand>-\u003Ctopic>-\u003Cvariable>-B`), then `set_timeline_document` on the new\n   id with the document `get_timeline` returned. The returned `timeline`\n   object is accepted as is, but its `name` renames the target sequence, so\n   set `name` to the variant's name before sending it.\n3. **Make the one change** on the copy with `edit_timeline`\n   (`timeline-edit-ops` owns the op shapes). A reorder is `move_clip`. A\n   different headline is the text clip's content. A cut beat is\n   `delete_clip`, then close the gap so the runtime stays matched.\n4. **A variant that needs new footage** gets its shot on the same board:\n   `edit_storyboard` `duplicate_shot` on the shot it replaces, `update_shot`\n   with the changed `action` or `motion` and `notes: \"variant B only\"`, then\n   `render_storyboard_clips {targets: [\u003Cslug>]}` for that shot alone, and place\n   the clip in the variant cut with `add_media_clip`. Assemble the control\n   before adding variant shots: `assemble_storyboard_timeline` lays every\n   rendered shot into the sequence, so re-running it afterwards would put the\n   variant shot into the control.\n5. **Prove the match.** `compare_timeline_frames {a: \u003Ccontrol id>, b:\n   \u003Cvariant id>}` sweeps both cuts. Frames should differ only where the\n   variable lives. A difference anywhere else is a second variable: fix it\n   before rendering.\n6. **Render each arm** with `render_timeline` and name the assets after the\n   timelines. Never overwrite a delivered control.\n\nNew renders cost money. Before rendering footage for a variant, report how\nmany clips the variants add and on which model, and get the user's go-ahead.\nA variant that only reorders, retimes or rewords existing material renders\nnothing new except the final `render_timeline`.\n\n## Check comprehension before anyone pays for media\n\nA variant that holds attention but blurs the message fails the test it was\nbuilt for. Before handing the arms over, ask a model that reads video what\neach one says: `understand_video` on each rendered file with \"What is being\nadvertised, what does it claim, and what is the viewer asked to do?\". It\ntakes a `provider` and `model`. A Gemini model reads the whole clip with its\naudio, and other vision models see sampled stills only. If an\narm's answer misses the central claim or the CTA, fix that arm first. This\nis a check on the cut, not a substitute for audience data.\n\n## The test plan\n\nDeliver the plan with the files. The user runs it on their platform.\n\n```text\nTEST:          \u003Cname>\nHYPOTHESIS:    \u003Cone sentence: change, expected effect, guard metric>\nVARIABLE:      \u003Cthe one difference between arms>\nARMS:          A = \u003Casset name> (control), B = \u003Casset name>\nHELD FIXED:    runtime, assets, music, CTA, audience, placement, budget, dates\nPRIMARY:       \u003Cthe metric the hypothesis names>\nGUARDRAILS:    \u003Ccomprehension or click-through, and conversion if tracked>\nSPLIT:         the platform's own split-test or experiment tool, so exposure\n               is randomized. Not two ads posted on different days\nSTOP RULE:     \u003Cequal spend per arm, and the platform's confidence reading,\n               before calling a winner>\n```\n\n## Reading the results\n\nRetention is a diagnostic, not the definition of success.\n\n- **A spike** in a retention curve can be rewatching or confusion. Check what\n  is on screen at that second: a dense frame people pause on, or a payoff\n  worth seeing again.\n- **A dip** can be people skipping ahead rather than leaving. Compare it with\n  the drop in viewers who reach the end.\n- **New and returning viewers** behave differently. YouTube reports them\n  apart. Compare like with like.\n- **Platform view metrics are not interchangeable.** TikTok reports 2-second,\n  6-second and 15-second views, and the definitions vary by report. Compare\n  the same metric across arms, from the same report.\n- **Completion without comprehension or conversion** means the ad entertained\n  the wrong audience. Weigh the primary metric against the guardrails before\n  naming a winner.\n- **A cliffhanger arm** that raises intent to watch the next episode has not\n  shown more enjoyment or more sales. Check what viewers did next, not only\n  what they said.\n\nWhen the arms ran at different times, to different audiences, or with\nunequal spend, say the result is confounded instead of naming a winner.\n\n## Keep what was learned\n\nA result is only useful if the next brief starts from it. Save the outcome\nwith `memory_save`: product, audience, placement, the variable, both arms'\nprimary and guardrail numbers, and the decision. Write the next test from\nthere.\n","# Testing ad structure with matched variants\n\nCreator frameworks for hooks and story loops have not been tested against\nalternatives in controlled comparisons. The way to know whether a structure\nworks for this product and this audience is to run two versions that differ\nin that structure and nothing else. This skill builds those versions in\nNodeTool, writes the plan the user runs on their ad platform, and reads what\ncomes back.\n\nNodeTool makes and renders the variants. It does not buy media or read ad\naccounts, so the user runs the test and brings the numbers back. Load\n`ad-story-loops` for the loop map the variants are built from, and the job\nskill that made the ad (`motion-ad`, `commercial-beat-sheet`,\n`launch-commercial`, `ugc-video`) for how its board and cut are built.\n\n## One variable per test\n\nA pair of variants answers one question. Hold everything else fixed:\n\n- runtime within 0.5s, the same aspect ratio and the same opening frame\n  unless the opening is the variable\n- the same product assets, cast, music bed, grade and type system\n- the same claims, the same CTA wording and the same end card\n- the same placement, audience, budget and dates on the platform\n\nChange the hook and the ending together and a win cannot be attributed to\neither. A third arm is fine when the user has the budget, but each arm still\ndiffers from the control in one way.\n\n## Variables worth testing\n\n| Test | Control | Variant | Primary outcome |\n|---|---|---|---|\n| Curiosity gap | States the result, then explains | Names the question, explains, then shows the result | Retention through the first payoff |\n| Earned surprise | The expected outcome | A coherent reversal set up earlier | Post-view satisfaction and recall |\n| Loop count | One central question | Several connected questions | Completion and comprehension |\n| Ending | Resolved | Open, pointing at the next episode | Stated intent to watch the next one, and actual continuation |\n| Person vs facts | Product facts alone | The same facts through a person's goal | Brand recall and qualified engagement |\n| Hook archetype | One archetype | Another, same body | 2s and 6s views, then retention |\n\nWrite each test as a hypothesis before building: \"Opening on the question\ninstead of the result keeps more viewers to the first payoff at 8s, without\nlowering click-through.\"\n\n## Build the variants\n\nBuild the control first and finish it. Every variant is a copy with one\nchange, never a second ad made from scratch.\n\n1. **Snapshot the control.** `create_timeline_version {timeline_id}` on the\n   finished cut, so the control can always be restored.\n2. **Copy the cut.** `get_timeline {timeline_id}`, then `create_timeline`\n   with the same `fps`, `width` and `height` and a versioned name\n   (`\u003Cbrand>-\u003Ctopic>-\u003Cvariable>-B`), then `set_timeline_document` on the new\n   id with the document `get_timeline` returned. The returned `timeline`\n   object is accepted as is, but its `name` renames the target sequence, so\n   set `name` to the variant's name before sending it.\n3. **Make the one change** on the copy with `edit_timeline`\n   (`timeline-edit-ops` owns the op shapes). A reorder is `move_clip`. A\n   different headline is the text clip's content. A cut beat is\n   `delete_clip`, then close the gap so the runtime stays matched.\n4. **A variant that needs new footage** gets its shot on the same board:\n   `edit_storyboard` `duplicate_shot` on the shot it replaces, `update_shot`\n   with the changed `action` or `motion` and `notes: \"variant B only\"`, then\n   `render_storyboard_clips {targets: [\u003Cslug>]}` for that shot alone, and place\n   the clip in the variant cut with `add_media_clip`. Assemble the control\n   before adding variant shots: `assemble_storyboard_timeline` lays every\n   rendered shot into the sequence, so re-running it afterwards would put the\n   variant shot into the control.\n5. **Prove the match.** `compare_timeline_frames {a: \u003Ccontrol id>, b:\n   \u003Cvariant id>}` sweeps both cuts. Frames should differ only where the\n   variable lives. A difference anywhere else is a second variable: fix it\n   before rendering.\n6. **Render each arm** with `render_timeline` and name the assets after the\n   timelines. Never overwrite a delivered control.\n\nNew renders cost money. Before rendering footage for a variant, report how\nmany clips the variants add and on which model, and get the user's go-ahead.\nA variant that only reorders, retimes or rewords existing material renders\nnothing new except the final `render_timeline`.\n\n## Check comprehension before anyone pays for media\n\nA variant that holds attention but blurs the message fails the test it was\nbuilt for. Before handing the arms over, ask a model that reads video what\neach one says: `understand_video` on each rendered file with \"What is being\nadvertised, what does it claim, and what is the viewer asked to do?\". It\ntakes a `provider` and `model`. A Gemini model reads the whole clip with its\naudio, and other vision models see sampled stills only. If an\narm's answer misses the central claim or the CTA, fix that arm first. This\nis a check on the cut, not a substitute for audience data.\n\n## The test plan\n\nDeliver the plan with the files. The user runs it on their platform.\n\n```text\nTEST:          \u003Cname>\nHYPOTHESIS:    \u003Cone sentence: change, expected effect, guard metric>\nVARIABLE:      \u003Cthe one difference between arms>\nARMS:          A = \u003Casset name> (control), B = \u003Casset name>\nHELD FIXED:    runtime, assets, music, CTA, audience, placement, budget, dates\nPRIMARY:       \u003Cthe metric the hypothesis names>\nGUARDRAILS:    \u003Ccomprehension or click-through, and conversion if tracked>\nSPLIT:         the platform's own split-test or experiment tool, so exposure\n               is randomized. Not two ads posted on different days\nSTOP RULE:     \u003Cequal spend per arm, and the platform's confidence reading,\n               before calling a winner>\n```\n\n## Reading the results\n\nRetention is a diagnostic, not the definition of success.\n\n- **A spike** in a retention curve can be rewatching or confusion. Check what\n  is on screen at that second: a dense frame people pause on, or a payoff\n  worth seeing again.\n- **A dip** can be people skipping ahead rather than leaving. Compare it with\n  the drop in viewers who reach the end.\n- **New and returning viewers** behave differently. YouTube reports them\n  apart. Compare like with like.\n- **Platform view metrics are not interchangeable.** TikTok reports 2-second,\n  6-second and 15-second views, and the definitions vary by report. Compare\n  the same metric across arms, from the same report.\n- **Completion without comprehension or conversion** means the ad entertained\n  the wrong audience. Weigh the primary metric against the guardrails before\n  naming a winner.\n- **A cliffhanger arm** that raises intent to watch the next episode has not\n  shown more enjoyment or more sales. Check what viewers did next, not only\n  what they said.\n\nWhen the arms ran at different times, to different audiences, or with\nunequal spend, say the result is confounded instead of naming a winner.\n\n## Keep what was learned\n\nA result is only useful if the next brief starts from it. Save the outcome\nwith `memory_save`: product, audience, placement, the variable, both arms'\nprimary and guardrail numbers, and the decision. Write the next test from\nthere.\n",{"name":5,"description":7},"main","c7226baa82daca42a19a5855c116544f97aa8bac","packages\u002Fsystem-skills\u002Fad-creative-testing",91,"https:\u002F\u002Fgithub.com\u002Fnodetool-ai\u002Fnodetool","https:\u002F\u002Fnodetool.ai\u002F","https:\u002F\u002Fgithub.com\u002Fnodetool-ai","No-code AI development","2026-11-15 03:27:49",{"source":8,"skillName":5,"commands":67,"prompt":77,"zipAvailable":78},[68,71,74],{"tool":69,"command":70},"npx","npx skills add nodetool-ai\u002Fnodetool --skill ad-creative-testing",{"tool":72,"command":73},"bunx","bunx skills add nodetool-ai\u002Fnodetool --skill ad-creative-testing",{"tool":75,"command":76},"pnpm","pnpm dlx skills add nodetool-ai\u002Fnodetool --skill ad-creative-testing","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fnodetool-ai\u002Fnodetool\u002Fblob\u002Fmain\u002Fpackages\u002Fsystem-skills\u002Fad-creative-testing\u002FSKILL.md",true,[],[81,90,125,151,166,180],{"skillNo":82,"name":83,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":85,"licenseFileUrl":6,"sourceHtmlUrl":86,"categories":87,"keywords":88,"publishedAt":89},"01M742MX3Y2ZYK884ZXTRFKB9P","humanizer","Rewrite AI-sounding text so it reads like the writer without changing what it says.\nUse when editing or reviewing prose for AI tells: not-X-but-Y contrasts, one-line\nclosers, staged openers, forced triads, dashes everywhere, inflated claims, sales\nlanguage, stock AI words, bold labels, or filler.","MIT","https:\u002F\u002Fgithub.com\u002Fnodetool-ai\u002Fnodetool\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fhumanizer\u002FSKILL.md",[],[],"2026-11-11 15:11:14",{"skillNo":91,"name":92,"nameZh":6,"description":93,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":94,"categories":95,"keywords":104,"publishedAt":52},"01M6BY4D93R0Q9400ZHJ04DXP1","nodetool","Use the NodeTool MCP server to build, validate, run, and debug visual AI workflows, and to work with assets, collections, and jobs. Use when the user mentions NodeTool or asks for a workflow graph, a node, or a media generation run on their NodeTool server.","https:\u002F\u002Fgithub.com\u002Fnodetool-ai\u002Fnodetool\u002Fblob\u002Fmain\u002Fplugins\u002Fchatgpt\u002Fnodetool\u002Fskills\u002Fnodetool\u002FSKILL.md",[96,99],{"categoryNo":97,"categoryName":98,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发",{"categoryNo":100,"categoryName":101,"taxonomyType":102,"taxonomyTypeDesc":103},"01JBSEEDCAT000000000000030","开发工程师","PROFESSION","适用职业",[92,105,106,107,33,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124],"MCP","Media","workflows","generation","media generation","with","visual","use","collections","for","node","work","debug","graph","validate","assets","run","server","mcp server","workflow",{"skillNo":126,"name":127,"nameZh":6,"description":128,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":129,"categories":130,"keywords":132,"publishedAt":52},"01M6BY4AEB05ZPBRVD078ATR2H","wan-2-6-prompting","Prompt Alibaba's Wan 2.6 video — text-to-video with a global style line plus timing-bracketed shots, and image-to-video that describes only what changes over time. Covers the shot-mode switch per provider, the prompt budget, negative prompts, motion intensity wording, audio input and prompt expansion. Use whenever the model id names a Wan 2.6 video route (alibaba\u002Fwan-2.6\u002Ftext-to-video, alibaba\u002Fwan-2.6\u002Fimage-to-video, wan-video\u002Fwan-2.6-t2v, wan-video\u002Fwan-2.6-i2v, wan-video\u002Fwan2.6-i2v-flash) on generate_video, animate_image or a TextToVideo node. Not for Wan 2.5 or 2.7, or the Wan 2.6 image routes.","https:\u002F\u002Fgithub.com\u002Fnodetool-ai\u002Fnodetool\u002Fblob\u002Fmain\u002Fpackages\u002Fsystem-skills\u002Fwan-2-6-prompting\u002FSKILL.md",[131],{"categoryNo":24,"categoryName":25,"taxonomyType":21,"taxonomyTypeDesc":22},[133,134,135,136,137,138,139,115,140,141,142,37,143,144,145,146,147,148,149,150,47],"prompting","wan 2 6 prompting","line","expansion","input","image-to-video","text-to-video","route","only","mode","flash","Budget","generate","text","image","animate","model","Audio",{"skillNo":152,"name":153,"nameZh":6,"description":154,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":155,"categories":156,"keywords":159,"publishedAt":52},"01M6BY47T4748ZG5ZMTAR97ZWZ","video-workflow","Build a reusable NodeTool video workflow for batches or repeated inputs. For a single finished piece, use a storyboard skill.","https:\u002F\u002Fgithub.com\u002Fnodetool-ai\u002Fnodetool\u002Fblob\u002Fmain\u002Fpackages\u002Fsystem-skills\u002Fvideo-workflow\u002FSKILL.md",[157,158],{"categoryNo":24,"categoryName":25,"taxonomyType":21,"taxonomyTypeDesc":22},{"categoryNo":97,"categoryName":98,"taxonomyType":21,"taxonomyTypeDesc":22},[124,160,47,161,162,163,164,112,165,92,114],"video workflow","reusable","storyboard","build","batches","single",{"skillNo":167,"name":168,"nameZh":6,"description":169,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":170,"categories":171,"keywords":173,"publishedAt":52},"01M6BY4587N2SYN315SXGEMZC3","video-clone","Analyze a reference video and recreate its shot structure with a new product or cast in NodeTool.","https:\u002F\u002Fgithub.com\u002Fnodetool-ai\u002Fnodetool\u002Fblob\u002Fmain\u002Fpackages\u002Fsystem-skills\u002Fvideo-clone\u002FSKILL.md",[172],{"categoryNo":24,"categoryName":25,"taxonomyType":21,"taxonomyTypeDesc":22},[174,175,47,110,176,177,92,178,49,38,179],"clone","video clone","cast","new","analyze","reference",{"skillNo":181,"name":182,"nameZh":6,"description":183,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":184,"categories":185,"keywords":187,"publishedAt":52},"01M6BY4257C2KQYYMD9EK6DK8E","video-audio-continuity","Keep sound continuous across a multi-scene piece cut from generated video — why one clip per scene hard-cuts the audio at every boundary, when to write all the scenes into a single generation instead, the sound-design vocabulary that survives a provider's copyright filter, and how to check the length, the aspect ratio and the mix of what came back. Use when a piece has more than one scene and the video model writes its own audio (Seedance 2, Veo 3, MiniMax H3, Kling 2.6 and later), or when a cut's sound drops out at the shot changes.","https:\u002F\u002Fgithub.com\u002Fnodetool-ai\u002Fnodetool\u002Fblob\u002Fmain\u002Fpackages\u002Fsystem-skills\u002Fvideo-audio-continuity\u002FSKILL.md",[186],{"categoryNo":24,"categoryName":25,"taxonomyType":21,"taxonomyTypeDesc":22},[188,47,150,189,190,191,39,192,193,194,195,196,197,33,198,199,149,200,51,201,202],"video audio continuity","continuity","write","why","veo 3","at","cut","later","across","filter","generated","how to","mix","boundary","how"]