Agent skill
shorts-cut
Select complete takes and remove false starts or dead air while preserving speech meaning and word boundaries.
From timscheuerai/gtm-architecture-template · 51 skill entries · 13 · pushed 2026-09-17
What it does when it runs
Select complete takes and remove false starts or dead air while preserving speech meaning and word boundaries.
Automated analysis of the skill and the 0 files bundled beside it. A skill’s own description is written to be selected by an agent, so it describes the job and not the dependencies.
- Keys and connectors you must supply
- None found.
- Hosts it reaches
- No third-party host appears in the skill or its bundled files.
- Tool permissions it declares
- No
allowed-toolsin the frontmatter. It only issues instructions, so there is nothing to bound. - Actions present in the files
- None. Instructions only.
Install it
View source on GitHub ↗git clone --depth 1 --filter=blob:none --sparse https://github.com/timscheuerai/gtm-architecture-template.git /tmp/gtm-architecture-template git -C /tmp/gtm-architecture-template sparse-checkout set ".agents/skills/shorts-cut" mkdir -p ~/.claude/skills/shorts-cut cp -R "/tmp/gtm-architecture-template/.agents/skills/shorts-cut/." ~/.claude/skills/shorts-cut/
Picked up without a restart. A project skill of the same name is shadowed by your personal one. For one repository only, swap ~/.claude/skills for .claude/skills. Claude Code docs ↗
The folder is the same in every client that implements the format — 46 of them — so if yours is not above, only the destination changes.
The skill
Source on GitHub ↗Reproduced in full from timscheuerai/gtm-architecture-template/blob/2c312c9398064ad3730c786647293458cb89632c/.agents/skills/shorts-cut/SKILL.md, which is licensed MIT (repository). 162 words, 1 heading.
Choose the usable performance
Read the full transcript and inspect/listen to candidate ranges. Determine whether repeats are retakes or intentional emphasis. The last complete attempt is a useful preference for retake-heavy footage; the strongest coherent take wins when the brief or performance supports it.
For each beat, record source, in/out times and the reason for choosing that range. Reject unfinished fragments even when they occur last. Retakes can happen inside a phrase; inspect at word level.
ASR timing is a starting point. Listen around every proposed cut so consonants and breath are not clipped. Remove unnecessary silence inside a take only when pacing improves; preserve intentional pauses.
Build the EDL from source time and derive output positions cumulatively. After changing the cut, regenerate downstream timing maps and captions. Do not hand-maintain two conflicting timelines.
Render or play every join in context. Return the selected ranges, rejected alternatives where useful, final duration and any uncertain transcription. Use shorts-qa for the completed output.
Need help setting it up?
This page tells you what shorts-cut does and what it needs. Cheetah builds the agent setup it runs inside: data, CRM, sequencing and the guardrails.
Book a call →The directory stays free. There is nothing gated behind this.