gooseworks-ai/goose-skills / skills/video/render-street-interview
render-street-interview
VerifiedWrite a street-interview ad from a complete inspected commercial interaction and current brand facts. Render product guessing, or prepare mic-only, prepared-sample or visible-task conversation script/prompt previews. Premise, actions and ad connection vary; capture rules stay fixed. Conversation media delivery is unverified.
brands/grimwater.json
View on GitHub{
"_comment": "The approved seed-4815 run, as data. This file plus format_spec.py reproduces that take's prompt byte for byte; selftest.py asserts it. Nothing here is format knowledge - every field is a creative choice this brand made.",
"brand": "Grimwater",
"slug": "gw",
"product": {
"noun": "can",
"phrase": "white can",
"reference_image": "assets/grimwater/reference-photos/grimwater-still-19oz-tallboy.png",
"appearance": "same white body, same dark blackletter lettering, same gold illustration",
"colour_ban": "It never turns gold or yellow.",
"opening_extra": " and no ring pull is touched",
"open_state": "open"
},
"location": {
"description": "a wide pavement on one street corner outside a red brick building wrapped in green scaffolding, with a bus shelter to the left, parked cars along the kerb and bare plane trees behind",
"landmarks": "The same scaffolding, bus shelter and parked cars"
},
"question": "Quick question. What's in this can?",
"props": "a carrier bag, a phone, a backpack",
"cast_size": 4,
"shots": [
{
"kind": "handover_first",
"subject": "a man in his fifties in a navy work coat with a carrier bag",
"pronoun": "he",
"noun": "man"
},
{
"kind": "react",
"noun": "man",
"reaction": "shrugs"
},
{
"kind": "handover",
"subject": "a heavyset balding man in a paint-stained work jacket",
"pronoun": "he"
},
{
"kind": "speak",
"noun": "man",
"reaction": "snorts",
"manner": "flatly",
"line": "Beer. Obviously."
},
{
"_comment": "The trailing comma on `subject` is deliberate and is used verbatim: the approved payload reads '...across her chest, on that same corner' while shot 3 reads '...work jacket on that same corner'. Kept so the approved prompt reproduces exactly.",
"kind": "handover",
"subject": "a woman in her thirties in teal medical scrubs with a bag strap across her chest,",
"pronoun": "she"
},
{
"kind": "speak",
"noun": "woman",
"reaction": "winces",
"manner": "",
"line": "Something that's gonna kill me."
},
{
"kind": "sip",
"subject": "A young woman in a yellow raincoat with a backpack"
},
{
"kind": "payoff",
"noun": "young woman",
"manner": "surprised",
"line": "Wait, that's water."
}
],
"generation": {
"seed": 4815,
"duration": 12,
"fast": false
},
"brand_layer": {
"_comment": "Everything drawn locally with PIL. No model renders a letter. `cuts` are the take's OWN internal cuts, MEASURED off seed 4815 with ffmpeg scene detection - never calculated from the script. Caption spans are MEASURED off the take's Whisper word timings (question 0.00-1.64, 5.30-6.64, 7.58-8.94, 10.74-11.90) and clamped inside the shot they belong to.",
"cuts": [
4.04,
6.71,
9.38
],
"ambience_gap": [
1.75,
2.15
],
"_ambience_gap_comment": "A speech-free window in THIS take, in seconds: start, duration. The ambience bed and the end card's room tone are both cut from it, so they are literally the same street on the same afternoon. It is per-take and must be measured: seed 4802's gap was at 7.6s, which on seed 4815 is the middle of 'something that's gonna kill me'. A bed built over speech loops that speech under the whole video, which is what the operator's 'random background noises' turned out to be.",
"series_header": "WHAT'S IN THE CAN?",
"logo": "assets/grimwater/logos/grimwater-logo-3840.png",
"end_card": [
"NOTHING TO FEAR.",
"IT'S ONLY WATER."
],
"captions": [
[
0.1,
2.15,
"WE ASKED PEOPLE|WHAT'S IN THIS CAN.",
"title",
"",
"We asked people what's in this can."
],
[
5.25,
6.68,
"BEER. OBVIOUSLY.",
"answer",
"PAINTER",
"Beer. Obviously."
],
[
7.5,
9.0,
"GONNA KILL ME.",
"answer",
"NURSE",
"Something that's gonna kill me."
],
[
10.7,
11.95,
"WAIT, THAT'S WATER.",
"payoff",
"STUDENT",
"Wait, that's water."
]
],
"grade": {
"black_lift": 0.024,
"saturation": 1.0,
"_comment": "SOLVED against gw-single-seed4815.mp4 by fit_grade.py, not a constant. Per-take: re-solve after any new generation. Verified black 8.78 (band 7.0-10.3), face_sat 0.239 (band 0.14-0.228)."
}
}
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |