agentleFS
perso-aiperso-ai/perso-dubbing-plugin/skills/clip/SKILL.md

clip

Skill40 starsChanged 4 months agoMITView on GitHub

Cut a video into short-form clips (highlights/shorts) — cut given timecodes directly, or pick the best moments from a Perso AI STT project. Reframes 16:9→9:16. Subtitles are added separately by the srt skill.

  • Reads credentials

13 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 0 are distinctive, and 4 are very specific to this project.

  • 0

    Common

    Most of the 10 edited versions we read keep these.

  • 0

    Distinctive

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Phase 2 — pick, then cut: Read the transcript and choose clips, each { "title": "...", "start_order": N, "end_order": M }.

    line 32
  2. Very specific:

    Run: node scripts/clip.mjs --project <seq> --clips '<json>' [--out folder] (--clips also accepts @path).

    line 43
  3. Very specific:

    Write the sidecars: node scripts/clip.mjs --sidecars "<out>/clips.json" — add --karaoke only for a karaoke preset (it also writes …

    line 49
  4. Very specific:

    Add --host <runtime> for telemetry (silent), same as the other workers.

    line 52

“Common” is measured: kept by at least 6 in 10 of the 10 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

/clip

Cut a video into short clips + reframe. Subtitles are not this skill's job — after cutting, the srt skill styles them (see the handoff below).

Entry points

  • Explicit timecodes ("cut 2:00–3:00 of this video") → cut the local file directly: --video "<file>" --ranges "2:00-3:00,...". No STT, no key, no credits.
  • AI highlight selection on an STT project → --plan then --clips (the flow below). Needs the project's transcript + word timestamps.
  • Raw video, "find the good moments" (no timecodes given) → you can't pick highlights without a transcript, so ask the user which they want:

    • STT (proper selection, uses credits): run the srt skill's STT first to create the project, then clip it with --plan/--clips.
    • Free (no STT): the user gives the timecodes and you cut them with --video --ranges.

    Proceed with their choice — don't run STT (which bills credits) without asking.

Core rules

  • Only the worker sees the raw key — never open, echo, or pass it as an argument.
  • The key gate applies only to --project/--plan/--clips. --video --ranges and --sidecars are fully offline — never run a key check/registration for them.
  • Run in the background — encoding takes time.
  • This skill shares the dubbing skill's libraries: the dubbing folder must be installed next to clip.

Flow

Phase 1 — plan. node scripts/clip.mjs --project <seq> --plan prints the summary's section map (topic guide) plus the full sentence transcript, each line as order [start-end] text.

Phase 2 — pick, then cut. Read the transcript and choose clips, each { "title": "...", "start_order": N, "end_order": M }. Rubric:

  • Hook first — start on a sentence that grabs attention (question, bold claim, funny/surprising line).
  • End where the moment resolves, not at the first complete sentence. Ride the beat to its end — the whole reaction, follow-through, and any secondary punchline. Read a few sentences past the apparent ending; keep them if still the same reaction.
  • Cut at the transition — end just before the energy drops or the tone shifts (excited reaction → calm narration).
  • Peak — favour funny / surprising / emotional / quotable moments; the high point often sits a sentence or two after the literal reveal.
  • Length 30–90s — the worker warns when outside; prefer the full arc over a premature cut.
  • Title — describes the moment; becomes the file name (01_<title>.mp4).

Sections are chapter-length — use the section map only as a guide, then tighten to a hook.

Run: node scripts/clip.mjs --project <seq> --clips '<json>' [--out folder] (--clips also accepts @path). Landscape (16:9) sources are reframed to 9:16 (drama-shorts letterbox); portrait sources are kept. The worker writes the clips plus a clips.json manifest (records each clip + the transcript slice needed for subtitles later). Relay [progress]; deliver the files from [clips-output].

Subtitles → hand off to the srt skill

Subtitles are off by default. After cutting, tell the user clips are ready and ask whether to add subtitles. If they want subtitles, show the presets first — never jump straight to one, exactly as the srt skill's styling flow does: the interactive gallery (select + Apply) in the Claude app, or the static ../srt/assets/style_presets.png + node ../srt/scripts/style.mjs --list-presets menu elsewhere. Let the user pick a preset or describe a style (short-form presets like karaoke/bold-punch suit clips). Only after they pick:

  1. Write the sidecars: node scripts/clip.mjs --sidecars "<out>/clips.json" — add --karaoke only for a karaoke preset (it also writes <clip>.timestamps.json for measured word timing; a plain preset needs just the .srt).
  2. Style each clip with the srt skill: node ../srt/scripts/style.mjs "<clip>.mp4" "<clip>.srt" --preset <id> (append --word-timestamps "<clip>.timestamps.json" for karaoke). See the srt skill's "Style & burn subtitles" section.

Add --host <runtime> for telemetry (silent), same as the other workers.

Config (env)

Same as /dubbing / /srt (PERSO_API_BASE/PERSO_MEDIA_BASE — https perso.ai hosts only, PERSO_SPACE_SEQ, XP_API_KEY, PERSO_NO_TELEMETRY, …).

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.