{"id":85187,"plugin_id":"plugin_asdk_app_6a18d4c190a0819186e6b129a09e931e","kind":"skill","collection_source":"plugin_package","comparison_source":"plugin_package","observed_at":"2026-11-07T00:09:12.840Z","digest":"b02e2ce8d27394cab28ea7aaab254b1b447e6c49b65d36ef2185fface826dcc4","against":65285,"payload":{"description":"Choose the moments of talking footage worth keeping and cut them so no one is clipped mid-sentence - the transcript sets every in and out, the watch picks what is worth it - then read the cut as the viewer will hear it with scripts/review_cuts.py, which flags chopped words, filler and seams that open on \"but\" or \"she\", with the time to move each edge to. Use when cutting a short, a reel or highlights from an interview, podcast, vlog, talk or any footage where people speak, and when a finished cut sounds choppy.","included_files":[{"relative_path":"scripts/review_cuts.py","size_in_bytes":23809}],"name":"speech-aware-cutting","skill_md_contents":"---\nname: speech-aware-cutting\ndescription: Choose the moments of talking footage worth keeping and cut them so no one is clipped mid-sentence - the transcript sets every in and out, the watch picks what is worth it - then read the cut as the viewer will hear it with scripts/review_cuts.py, which flags chopped words, filler and seams that open on \"but\" or \"she\", with the time to move each edge to. Use when cutting a short, a reel or highlights from an interview, podcast, vlog, talk or any footage where people speak, and when a finished cut sounds choppy.\n---\n\n# Speech cuts\n\nA short cut from talking footage lives on its words. The framing can be perfect and the short is\nstill unwatchable when it opens on half a sentence or spends its first two seconds on \"yeah, no,\ntotally\". So the cut is planned from what is **said**, and checked as what is **heard**, before\nanything is built.\n\n## Read the footage\n\nRead every source as the footage-analysis skill says for talking footage: a transcript kept as the\ntool returned it (`transcript-<source id>.json` where you run commands), which sets every boundary, and a watch, which picks the moments. A photo has no speech\nat all: it is a beat, not a backbone.\n\n## Plan the cut\n\n1. **One pool of moments.** Take the watch's candidate moments from every source into one list,\n   each with the source it came from: its times count on that source's clock, as the footage-reframing\n   skill says.\n2. **Snap each to whole thoughts.** A clip's `in` is a sentence's first word, or a word after a\n   real pause; its `out` is a sentence's last word, or a word before one. The transcript's\n   `cutPoints` mark the clean ends.\n3. **Order and trim to the target.** Order the moments by their loops, as the retention-hooks\n   skill says. Aim for the target length within about 15%. **When the length and a clean sentence\n   conflict, the sentence wins and the length moves.** Never cut a word out of a sentence to save time: the ear hears the\n   splice at once, where the eye would forgive a picture cut.\n4. **Place photos at sentence edges**, never inside one, each held long enough to read.\n5. **Keep what the user saw work.** A recut or a fix starts from the last version the user saw,\n   never from zero. Every moment they liked stays, with the picture it had; a fix for a constraint\n   (a clipped word, a pause, the length) moves an edge, never swaps the moment for another. A\n   moment dropped, moved or given a new picture gets one line in the message to the user. Where\n   the use case keeps the edit-pipeline record, its keepers ledger and `scripts/meaning_diff.py`\n   do this, as the edit-pipeline skill says.\n\nWrite the plan as the reel's clip list, the one the footage-reframing skill's `scripts/reel_cut.py` reads,\nwith each spoken source's `transcript` beside its other fields (its file path, where you run commands).\n\n## Review it as heard\n\nRun `python3 scripts/review_cuts.py reel.json` on that same file. It joins each clip's words in\nplay order into the **heard script**, and checks three things per clip:\n\n| Check | Fails when | Its repair |\n|---|---|---|\n| Integrity | an edge cuts through a word, or the clip opens or stops mid-sentence | move that edge to the time given |\n| Substance | most of the clip is filler; the opening clip is held to a stricter share | drop it, or take the sentence beside it |\n| Fluency | a clip opens on a connective (\"but\", \"so\") with nothing before it, or the reel opens on a pronoun no one has been named for | move the in point back to what it refers to, or lead with another clip |\n| Context | a clip opens on an answer (\"Yes\", \"Exactly\") whose question the cut drops | move the in point back to the question |\n\nEach clip also lists the `speakers` its words carry. Use them when the footage-reframing skill asks who a\nclip should hold on.\n\nThe review also returns `context`: each name the reel says for the first time, each \"she\" or \"that\"\nright after a seam, each \"as I said\", and each seam that jumps far in its source, with the second of\nthe reel it is heard at and, when the cut drops it, what the source said first. These are not faults.\nThey are the places a listener who knows only the reel may get lost.\n\n**Then have the script read cold.** You know the whole footage, so you cannot hear what is missing.\nHave it read on paper as the quality-review skill's cold watch says, its calls quoted first as it\nsays: a SPEND (quote it in one line, then get the yes as an ASK with \"Yes, go ahead\" / \"Yes, up to N credits for this task\" / \"No\": the host's own choice tool where it offers one, otherwise those options numbered in one message; the host's own confirmation before a tool runs is not this yes; unless a spend limit the user stated covers it: N credits or an explicit no-limit (\"as much as needed\"), never inferred from words like \"go ahead\". Within it, ask nothing: quote it in one line (\"About 8 credits; 12 of 20 left after this\"), call it, then say what it spent and what remains, keeping a running total that includes failed calls, and stop to ask before the first call that would cross it. A money amount is asked once to restate it in credits; in unattended pace, never ask: call only with a consent or a stated spend limit already recorded, otherwise write the quote to pendingSpends and stop, as picsart-video-rules says) before the first read unless every call is quoted at 0 credits. Each row a clip's time in the\nreel and its heard words, so readers who know nothing read it up to each moment and say what they\nunderstood. Check their account at each `context` time: did they know who or what it was? Where\nthey did not, keep the setup or lead with another clip. Read the `script` yourself too, as prose,\nfor what is clunky; whether a clean clip is worth its seconds is your call.\n\nApply the repairs to the plan and run it again until it is clean. A repair changes numbers in a\nplan, so it costs nothing. **The review never blocks the cut:** if every candidate is thin, build\nfrom the strongest and say so in one line.\n\nA clip it marks `silent` (a photo, a shot with no transcript) is a beat of silence in the heard\nscript. A segment the transcript returned without word times is checked only at its sentence\nedges, and the review says so.\n"},"changes":[{"path":"/included_files","type":"changed","before":[{"relative_path":"scripts/review_cuts.py","size_in_bytes":15125}],"after":[{"relative_path":"scripts/review_cuts.py","size_in_bytes":23809}]},{"path":"/skill_md_contents","type":"changed","before":"---\nname: speech-aware-cutting\ndescription: Choose the moments of talking footage worth keeping and cut them so no one is clipped mid-sentence - the transcript sets every in and out, the watch picks what is worth it - then read the cut as the viewer will hear it with scripts/review_cuts.py, which flags chopped words, filler and seams that open on \"but\" or \"she\", with the time to move each edge to. Use when cutting a short, a reel or highlights from an interview, podcast, vlog, talk or any footage where people speak, and when a finished cut sounds choppy.\n---\n\n# Speech cuts\n\nA short cut from talking footage lives on its words. The framing can be perfect and the short is\nstill unwatchable when it opens on half a sentence or spends its first two seconds on \"yeah, no,\ntotally\". So the cut is planned from what is **said**, and checked as what is **heard**, before\nanything is built.\n\n## Read the footage\n\nRead every source as the footage-analysis skill says for talking footage: a transcript saved as\n`transcript-<source id>.json`, which sets every boundary, and a watch, which picks the moments. A photo has no speech\nat all: it is a beat, not a backbone.\n\n## Plan the cut\n\n1. **One pool of moments.** Take the watch's candidate moments from every source into one list,\n   each with the source it came from: its times count on that source's clock, as the footage-reframing\n   skill says.\n2. **Snap each to whole thoughts.** A clip's `in` is a sentence's first word, or a word after a\n   real pause; its `out` is a sentence's last word, or a word before one. The transcript's\n   `cutPoints` mark the clean ends.\n3. **Order and trim to the target.** Order the moments by their loops, as the retention-hooks\n   skill says. Aim for the target length within about 15%. **When the length and a clean sentence\n   conflict, the sentence wins and the length moves.** Never cut a word out of a sentence to save time: the ear hears the\n   splice at once, where the eye would forgive a picture cut.\n4. **Place photos at sentence edges**, never inside one, each held long enough to read.\n\nWrite the plan as the reel's clip list, the one the footage-reframing skill's `scripts/reel_cut.py` reads,\nwith each spoken source's `transcript` path beside its other fields.\n\n## Review it as heard\n\nRun `python3 scripts/review_cuts.py reel.json` on that same file. It joins each clip's words in\nplay order into the **heard script**, and checks three things per clip:\n\n| Check | Fails when | Its repair |\n|---|---|---|\n| Integrity | an edge cuts through a word, or the clip opens or stops mid-sentence | move that edge to the time given |\n| Substance | most of the clip is filler; the opening clip is held to a stricter share | drop it, or take the sentence beside it |\n| Fluency | a clip opens on a connective (\"but\", \"so\") with nothing before it, or the reel opens on a pronoun no one has been named for | move the in point back to what it refers to, or lead with another clip |\n\nEach clip also lists the `speakers` its words carry. Use them when the footage-reframing skill asks who a\nclip should hold on.\n\n**Then read the `script` yourself, as prose.** If it does not hold together as a paragraph, it\nwill not hold together as a video. The script finds the mechanical faults; whether a clean clip is\nworth its seconds is your call.\n\nApply the repairs to the plan and run it again until it is clean. A repair changes numbers in a\nplan, so it costs nothing. **The review never blocks the cut:** if every candidate is thin, build\nfrom the strongest and say so in one line.\n\nA clip it marks `silent` (a photo, a shot with no transcript) is a beat of silence in the heard\nscript. A segment the transcript returned without word times is checked only at its sentence\nedges, and the review says so.\n","after":"---\nname: speech-aware-cutting\ndescription: Choose the moments of talking footage worth keeping and cut them so no one is clipped mid-sentence - the transcript sets every in and out, the watch picks what is worth it - then read the cut as the viewer will hear it with scripts/review_cuts.py, which flags chopped words, filler and seams that open on \"but\" or \"she\", with the time to move each edge to. Use when cutting a short, a reel or highlights from an interview, podcast, vlog, talk or any footage where people speak, and when a finished cut sounds choppy.\n---\n\n# Speech cuts\n\nA short cut from talking footage lives on its words. The framing can be perfect and the short is\nstill unwatchable when it opens on half a sentence or spends its first two seconds on \"yeah, no,\ntotally\". So the cut is planned from what is **said**, and checked as what is **heard**, before\nanything is built.\n\n## Read the footage\n\nRead every source as the footage-analysis skill says for talking footage: a transcript kept as the\ntool returned it (`transcript-<source id>.json` where you run commands), which sets every boundary, and a watch, which picks the moments. A photo has no speech\nat all: it is a beat, not a backbone.\n\n## Plan the cut\n\n1. **One pool of moments.** Take the watch's candidate moments from every source into one list,\n   each with the source it came from: its times count on that source's clock, as the footage-reframing\n   skill says.\n2. **Snap each to whole thoughts.** A clip's `in` is a sentence's first word, or a word after a\n   real pause; its `out` is a sentence's last word, or a word before one. The transcript's\n   `cutPoints` mark the clean ends.\n3. **Order and trim to the target.** Order the moments by their loops, as the retention-hooks\n   skill says. Aim for the target length within about 15%. **When the length and a clean sentence\n   conflict, the sentence wins and the length moves.** Never cut a word out of a sentence to save time: the ear hears the\n   splice at once, where the eye would forgive a picture cut.\n4. **Place photos at sentence edges**, never inside one, each held long enough to read.\n5. **Keep what the user saw work.** A recut or a fix starts from the last version the user saw,\n   never from zero. Every moment they liked stays, with the picture it had; a fix for a constraint\n   (a clipped word, a pause, the length) moves an edge, never swaps the moment for another. A\n   moment dropped, moved or given a new picture gets one line in the message to the user. Where\n   the use case keeps the edit-pipeline record, its keepers ledger and `scripts/meaning_diff.py`\n   do this, as the edit-pipeline skill says.\n\nWrite the plan as the reel's clip list, the one the footage-reframing skill's `scripts/reel_cut.py` reads,\nwith each spoken source's `transcript` beside its other fields (its file path, where you run commands).\n\n## Review it as heard\n\nRun `python3 scripts/review_cuts.py reel.json` on that same file. It joins each clip's words in\nplay order into the **heard script**, and checks three things per clip:\n\n| Check | Fails when | Its repair |\n|---|---|---|\n| Integrity | an edge cuts through a word, or the clip opens or stops mid-sentence | move that edge to the time given |\n| Substance | most of the clip is filler; the opening clip is held to a stricter share | drop it, or take the sentence beside it |\n| Fluency | a clip opens on a connective (\"but\", \"so\") with nothing before it, or the reel opens on a pronoun no one has been named for | move the in point back to what it refers to, or lead with another clip |\n| Context | a clip opens on an answer (\"Yes\", \"Exactly\") whose question the cut drops | move the in point back to the question |\n\nEach clip also lists the `speakers` its words carry. Use them when the footage-reframing skill asks who a\nclip should hold on.\n\nThe review also returns `context`: each name the reel says for the first time, each \"she\" or \"that\"\nright after a seam, each \"as I said\", and each seam that jumps far in its source, with the second of\nthe reel it is heard at and, when the cut drops it, what the source said first. These are not faults.\nThey are the places a listener who knows only the reel may get lost.\n\n**Then have the script read cold.** You know the whole footage, so you cannot hear what is missing.\nHave it read on paper as the quality-review skill's cold watch says, its calls quoted first as it\nsays: a SPEND (quote it in one line, then get the yes as an ASK with \"Yes, go ahead\" / \"Yes, up to N credits for this task\" / \"No\": the host's own choice tool where it offers one, otherwise those options numbered in one message; the host's own confirmation before a tool runs is not this yes; unless a spend limit the user stated covers it: N credits or an explicit no-limit (\"as much as needed\"), never inferred from words like \"go ahead\". Within it, ask nothing: quote it in one line (\"About 8 credits; 12 of 20 left after this\"), call it, then say what it spent and what remains, keeping a running total that includes failed calls, and stop to ask before the first call that would cross it. A money amount is asked once to restate it in credits; in unattended pace, never ask: call only with a consent or a stated spend limit already recorded, otherwise write the quote to pendingSpends and stop, as picsart-video-rules says) before the first read unless every call is quoted at 0 credits. Each row a clip's time in the\nreel and its heard words, so readers who know nothing read it up to each moment and say what they\nunderstood. Check their account at each `context` time: did they know who or what it was? Where\nthey did not, keep the setup or lead with another clip. Read the `script` yourself too, as prose,\nfor what is clunky; whether a clean clip is worth its seconds is your call.\n\nApply the repairs to the plan and run it again until it is clean. A repair changes numbers in a\nplan, so it costs nothing. **The review never blocks the cut:** if every candidate is thin, build\nfrom the strongest and say so in one line.\n\nA clip it marks `silent` (a photo, a shot with no transcript) is a beat of silence in the heard\nscript. A segment the transcript returned without word times is checked only at its sentence\nedges, and the review says so.\n"}],"summary":"Fields changed: 2. /included_files, /skill_md_contents.","summary_kind":"deterministic","summary_metadata":{}}