{"id":42783,"plugin_id":"plugin_asdk_app_6a8c52a06c3081919407e081a375c210","kind":"skill","collection_source":"plugin_package","comparison_source":null,"observed_at":"2026-10-24T13:07:06.019Z","digest":"ee8be63448337447c5c2ba98fd0087b3fc0461c2edc952517dd72ba2ec213b48","against":null,"payload":{"description":"Create one cinematic VFX shot with one hero centered inside a busy crowd: the hero snaps their fingers, every surrounding person and environmental action freezes while the hero remains active, then a second finger snap resumes the exact interrupted motion; use for stop time, frozen crowd, finger-snap time control, tempo fermo, folla congelata, detener el tiempo, or multitud congelada, not for bullet-time orbits, generic slow motion, static tableaux, or freeze-frame transitions.","included_files":[],"name":"vfx-stop-the-time","skill_md_contents":"---\nname: vfx-stop-the-time\ndescription: \"Create one cinematic VFX shot with one hero centered inside a busy crowd: the hero snaps their fingers, every surrounding person and environmental action freezes while the hero remains active, then a second finger snap resumes the exact interrupted motion; use for stop time, frozen crowd, finger-snap time control, tempo fermo, folla congelata, detener el tiempo, or multitud congelada, not for bullet-time orbits, generic slow motion, static tableaux, or freeze-frame transitions.\"\n---\n\n# Stop the Time VFX\n\n## Tool requirements\n\n- **Core evidence and source binding:** `prostudio_brief_get`, `prostudio_workspace_get`, `prostudio_references_list`, `prostudio_references_resolve`, `get_prostudio_media_catalog`, and `inspect_prostudio_media`.\n- **Final Stop the Time clip:** `prostudio_models_list`, `prostudio_generation_prepare`, `prostudio_video_generate`, `prostudio_job_status`, and `prostudio_generation_get`.\n- This declaration does not activate tools. Use only the tools required by the selected workflow. If a required tool is absent from the current scoped catalog and `load_prostudio_tools` is available, call it with the missing outcome before claiming the capability is unavailable. If retrieval still exposes no matching tool, report that limitation. A model-catalog lookup alone is never completion for a requested Stop the Time shot.\n\nCreate one continuous cinematic VFX clip centered on one recognizable hero inside a crowded, visibly active environment. The hero performs one clearly readable finger snap; every surrounding person, moving object, and environmental action freezes at the same instant while the hero remains alive and in control. The hero then performs a second readable finger snap and the entire world resumes from the exact interrupted state and trajectory.\n\nThis is a **causal double-snap time stop**, not a bullet-time orbit, generic slow motion, editing freeze-frame, or crowd tableau. The two snaps must visibly cause the stop and restart. The main subject is the only time-immune person unless the user explicitly identifies another exception.\n\nUse semantic skill selection only. Do not add deterministic routing, keyword matching, catalog changes, or unrelated UI changes as part of this workflow.\n\n## Minimum input and source policy\n\n- One clear reference of the intended hero is sufficient to begin. Preserve face, body proportions, hair, wardrobe, accessories, handedness when visible, and defining features.\n- Do not require separate references for the crowd, location, props, or background by default. Infer a coherent busy environment from the request or optional Brief and populate it with varied supporting people whose actions make the freeze readable.\n- An optional source image or video may define the setting. Load `media-inspection` and inspect every selected source before planning; filenames, thumbnails, prompts, and summaries are not visual evidence.\n- Load `reference-management` and bind the hero reference explicitly for identity and performance. Bind an optional environment source for location, composition, lighting, or motion only when the user supplies it.\n- If the user supplies a video, preserve its recognizable environment, camera language, crowd choreography, hero performance, lighting, and temporal cadence outside the localized time-stop window. Apply the effect at the most readable moment rather than remaking the whole clip.\n- If the user supplies an image, reconstruct a complete moving scene with coherent foreground, hero, midground crowd, background activity, ground contact, depth, occlusion, and lighting. The image must become a living scene before the first snap and after the second; do not animate only the camera over a flat card.\n- If no environment is supplied, choose a location where simultaneous activity reads immediately, such as a station concourse, city crossing, market, plaza, lobby, party, café, or event space. Keep the location subordinate to the hero and consistent with the requested tone.\n- Ask one concise question only when the intended hero is ambiguous or the setting materially changes the concept. Otherwise infer non-critical details and proceed. Never require a Brief to run this skill.\n\n## Stage the hero and crowd\n\n- Keep one hero visually centered or strongly center-weighted throughout the shot. The crowd surrounds the hero in foreground, midground, and background layers without blocking the face or snapping hand at either trigger.\n- Establish a genuinely busy world before the effect: people walk in different directions, talk, gesture, carry objects, cross paths, serve food, check phones, move luggage, cycle, or perform other location-appropriate actions. Include several distinct motion vectors and depth planes so the global freeze is undeniable.\n- Avoid a wall of duplicated extras. Supporting people must vary in age range, silhouette, wardrobe, direction, pace, pose, and task. Do not clone faces, bodies, outfits, props, or gait cycles.\n- Give the hero a calm, intentional performance. Before the first snap, the hero notices or controls the surrounding activity; during the freeze, the hero may breathe, blink, shift gaze, lower the snapping hand, or take one restrained step while everyone else remains perfectly suspended; before the second snap, the hand returns to a clear readable position.\n- Keep both snapping gestures anatomically credible. Show a visible preparation, thumb-to-finger contact, release, and small hand recoil. Preserve five fingers, stable wrist anatomy, consistent hand identity, and the same chosen snapping hand unless the user requests otherwise.\n- Frame the snapping hand clearly enough to read without turning the shot into an isolated hand close-up. Do not hide either snap behind a person, prop, motion blur, cut, flash, smoke, or camera whip.\n\n## Build the causal time-stop effect\n\nAuthor the result as one continuous shot with five ordered phases. Rebalance timings for the chosen duration, but never remove either snap or the resumed action:\n\n1. **Living baseline:** establish the centered hero and a crowded environment in active motion. Crowd actions, trajectories, ambience, and spatial depth must be readable before anything stops.\n2. **First snap and stop wave:** the hero performs the first complete finger snap. A subtle, motivated stop cue may radiate through sound, air, light, or restrained pressure distortion, but it must not obscure the scene. All non-hero motion locks at one shared instant immediately after the snap.\n3. **Frozen world:** the hero remains alive while every surrounding person and environmental action holds perfectly. Preserve mid-step balance, cloth shape, hair, droplets, steam, dust, vehicles, screens, reflections, shadows, and object ownership at the frozen instant. A controlled camera translation may reveal parallax and confirm real spatial depth, but never perform a full 360-degree orbit.\n4. **Second snap and release:** the hero performs a second complete finger snap. The same causal cue releases the world; no one anticipates the release, changes pose early, or resumes before the gesture completes.\n5. **Living resolution:** every person, object, and environmental motion continues from the exact frozen position, direction, velocity context, and action phase. Let the interrupted actions resolve long enough to prove continuity.\n\nFor an 8-second clip, a strong default is approximately 0.0–2.0 seconds living baseline, 2.0–2.5 seconds first snap and stop, 2.5–5.2 seconds frozen world, 5.2–5.7 seconds second snap and release, and 5.7–8.0 seconds living resolution.\n\n## Freeze continuity rules\n\n- Freeze the whole surrounding world, not only the crowd. People, vehicles, animals, falling objects, liquid, smoke, steam, dust, loose paper, foliage, fabric, hair, practical mechanisms, screen animation, and location-appropriate motion all stop together.\n- The hero remains temporally active. Hero breathing, eye movement, intentional head movement, hand motion, and one restrained repositioning may continue, while wind-driven hero hair or clothing should become quiet enough not to imply that ambient time is still running.\n- Preserve physical contacts. Feet remain planted or suspended in credible mid-step balance; hands stay attached to carried objects; cups, phones, bags, tools, and luggage do not drift; shadows and reflections remain attached to their causes.\n- The frozen crowd must not become statues, mannequins, blurred ghosts, duplicated cutouts, or background stills. Preserve living skin, natural expressions, fabric detail, depth, and exact interrupted poses.\n- On release, continue every trajectory forward. Do not reverse motion, replay the baseline, jump to new poses, swap extras, change wardrobe, teleport objects, or restart crowd loops from their beginning.\n- Keep the location, lighting, exposure, weather, crowd density, screen direction, and camera continuity stable across both snaps. The only global state change is motion running → motion stopped → motion running.\n\n## Camera, light, and sound\n\n- Use one controlled continuous camera path that keeps the hero center-weighted and both snapping hands readable. Favor a slow dolly, restrained lateral arc, or gentle push that creates foreground and background parallax without distracting from the causal gesture.\n- During the frozen phase, use modest camera translation to prove that the scene is volumetric and everyone is truly suspended. Do not use a full orbit, camera roll, whip pan, crash zoom, drone spiral, or a camera move so large that the hero leaves the center of the composition.\n- Choose a lens and distance that preserve the hero while showing enough crowd behavior to prove the effect. A 28–40 mm full-frame-equivalent lens, eye-to-chest camera height, moderate depth of field, controlled focus, and natural perspective are strong defaults.\n- Preserve realistic motion blur during the living phases. At the frozen instant, the held people and objects retain a crisp, physically plausible pose rather than smeared motion trails; the moving camera may retain subtle background parallax blur when appropriate.\n- Use motivated cinematic lighting with coherent contact shadows, reflections, and exposure. A restrained pulse or pressure ripple may mark each snap, but avoid magical beams, explosions, portals, color-world replacements, or large flashes unless explicitly requested.\n- Make both snaps audible when the selected model supports audio. A strong treatment is busy synchronized ambience before the first snap, one precise dry snap and a rapid ambience cutoff, a restrained suspended room tone while the hero remains active, a second precise snap, then the exact ambience and interrupted sounds resuming at normal cadence.\n- Audio must obey the same exception rule as the image: hero breathing, footsteps, or clothing may remain subtly audible during the freeze, while crowd voices, footsteps, traffic, machines, and environmental motion stop together.\n\n## Prompt strategy\n\n- Load `generation-planning`, inspect the live model catalog, and select a currently available reference-compatible video mode. Do not hardcode a provider or promise deterministic identity, hand anatomy, exact crowd count, or unsupported source conditioning.\n- Prepare exactly one final video generation and execute exactly one `prostudio_video_generate` action. Do not generate the five phases separately and do not assemble the effect in the Editor.\n- Write one spatially explicit English prompt naming the bound hero, environment, crowd layers and actions, snapping hand, first and second snap, frozen-world exception rule, timing, camera path, lens, lighting, aspect ratio, motion blur, audio behavior, and exact release continuity.\n- Use this structure and replace every bracketed field with concrete information:\n\n  ```text\n  ONE CONTINUOUS HIGH-END CINEMATIC DOUBLE-SNAP TIME-STOP VFX SHOT, one centered hero, one busy crowd, one complete stop-and-resume event, no cuts.\n  HERO LOCK: use [bound reference] as the exact identity and performance anchor; preserve [face, body, hair, wardrobe, accessories, and snapping hand]. The hero alone remains active while surrounding time is stopped.\n  ENVIRONMENT AND CROWD: [specific crowded location] with layered foreground, midground, and background activity: [distinct people, actions, props, vehicles, particles, and environmental motion]. Supporting people are varied and non-duplicated.\n  PHASE 1 [start–end] LIVING BASELINE: [hero position and behavior]; the crowd performs [readable simultaneous actions] with coherent trajectories, depth, contact shadows, and synchronized ambience.\n  PHASE 2 [start–end] FIRST SNAP / STOP: the hero visibly prepares and completes one anatomically correct finger snap with [chosen hand]. Immediately after contact and release, every non-hero person, object, particle, reflection, shadow-linked motion, vehicle, and environmental action freezes at one shared instant.\n  PHASE 3 [start–end] FROZEN WORLD: the hero remains alive and [small intentional performance] while the entire surrounding crowd and environment hold perfectly in their interrupted poses. The camera makes [restrained translating move] with real parallax and stable depth; no full orbit. Preserve every contact, trajectory context, identity, and frozen expression.\n  PHASE 4 [start–end] SECOND SNAP / RELEASE: the hero clearly completes a second anatomically correct snap with the same hand. Only after the snap completes does the surrounding world resume.\n  PHASE 5 [start–end] LIVING RESOLUTION: every interrupted action continues forward from the exact frozen pose, position, direction, and timing; the crowd never resets, reverses, swaps, or jumps.\n  CAMERA: [capture intent, focal length, height, framing, controlled path, focus, depth of field, frame rate, shutter and motion blur, level horizon, aspect ratio]; keep the hero center-weighted and both snaps unobstructed.\n  LIGHT, GRADE, AND AUDIO: [motivated source direction, exposure, contrast, atmosphere, and color intent]; busy ambience → first dry snap and synchronized silence → suspended hero-only detail → second dry snap → exact ambience and action resume.\n  CONTINUITY: one hero, one crowd, one location, one continuous camera path, two readable snaps, one exact frozen state, and one causal return to motion.\n  ```\n- Put positive behavior before constraints. Then explicitly forbid a 360-degree bullet-time orbit, camera roll, generic slow motion, hard cut, freeze-frame edit, static crowd from the first frame, hero freezing with the crowd, crowd members moving during the hold, early release, pose reset, reversed motion, looped crowd, duplicate extras, identity drift, changed wardrobe, broken hands, extra fingers, hidden snaps, floating props, foot sliding, mannequin skin, melting anatomy, exposure flash, portal, smoke wipe, scene replacement, and unrelated magical effects.\n\n## Prepare, generate, and inspect\n\n- Before generation, verify one exact hero reference is bound; the busy location and crowd actions are explicit; the hero is center-weighted; the chosen snapping hand and both gestures are readable; the hero-only exception is explicit; the five phases fit the duration; and the live model supports the selected binding, format, and audio behavior.\n- Prepare the exact generation and report the returned model, bound hero, duration, aspect ratio, audio behavior, and credit charge, then stop for the host's required confirmation. After approval, execute the prepared action without changing its arguments.\n- Poll the durable job and inspect the completed video before calling the result successful. A completed job is not proof that both snaps are readable or that the crowd froze and resumed coherently.\n- Inspect the living baseline, first hand preparation and snap, first frame of the freeze, middle and end of the hold, second hand preparation and snap, first resumed frames, and final resolution. Check hero identity, hand anatomy, crowd diversity, freeze simultaneity, frozen contacts, camera parallax, lighting, shadows, reflections, motion blur, audio cutoff, and trajectory continuity.\n\n## Preflight checklist\n\nBefore generation, verify that:\n\n- one recognizable hero is the sole time-immune subject and remains center-weighted;\n- the scene is visibly crowded and alive before the first snap, with distinct actions across multiple depth planes;\n- the first and second snaps are complete, anatomically plausible, unobstructed, and assigned to one consistent hand;\n- the first snap causes one synchronized global stop around the hero;\n- the frozen phase holds every non-hero person, object, particle, vehicle, reflection-linked motion, and environmental action consistently;\n- the camera creates modest spatial parallax without becoming a 360-degree orbit;\n- the second snap alone causes the release;\n- every interrupted action resumes forward from the exact frozen state without reset, reversal, swap, or teleport;\n- image-derived scenes are reconstructed as complete moving environments and video-derived scenes preserve their reference fidelity outside the localized effect;\n- one final generation contains the full baseline → first snap → freeze → second snap → resume sequence;\n- audio behavior supports both causal gestures and the global stop;\n- the live model supports the exact bindings and requested output settings.\n\n## Review and revision criteria\n\n- **Pass:** one centered hero performs two clearly readable snaps; the first freezes a previously active surrounding crowd and environment while the hero remains alive, and the second resumes every interrupted action from the exact held state.\n- **Pass:** the crowd contains varied people and tasks across real depth planes; frozen bodies, hands, feet, props, particles, shadows, reflections, and environmental motion remain coherent throughout the hold.\n- **Pass:** the camera proves volumetric depth with controlled translation and parallax while preserving the hero and both gestures; the effect reads without a full 360-degree orbit, cut, flash, or scene replacement.\n- **Pass:** identity, wardrobe, hand anatomy, lighting, exposure, crowd density, location geography, sound, and action trajectories remain continuous across both snaps.\n- **Fail:** the scene begins already frozen, only the hero or only part of the crowd stops, the hero freezes too, the crowd drifts during the hold, or the result is merely slow motion.\n- **Fail:** either snap is hidden, anatomically broken, incomplete, or disconnected from the state change; the release begins early or occurs without the second snap.\n- **Fail:** people or props reset, reverse, teleport, swap identity, duplicate, melt, slide, lose contact, or resume from a different pose or trajectory.\n- **Fail:** the camera performs bullet time, a complete orbit, lens-axis roll, digital spin, whip, zoom-only move, hard cut, transition wipe, or another effect that competes with the double-snap causal story.\n- For a localized failure, revise only the crowd action staging, hand visibility, snap timing, freeze simultaneity, hero-only exception, camera displacement, release continuity, or audio cue, then regenerate the complete single clip. Do not patch separate phases in the Editor.\n"},"changes":[],"summary":"First saved snapshot. No earlier version is available for comparison.","summary_kind":"deterministic","summary_metadata":{}}